Beginner is stuck with SPSS
Locked
 
|
Hi all! I’m stuck with my report. I collected some data, but now I don’t know what to do in SPSS. First some background about my report and data collection: I interviewed 30 people. They had to choose between 6 ‘cards’. Every card represent a product. This way I find out which product my participants would buy. If they choose one ‘card’, I asked two questions. One question about credibility and one about where to buy this product. Both with a Likert scale of 1 to 5. I did this until there were no ‘cards’ left. Thus the results are something like this for example: Card 5, 3, 4; Card 1, 2, 2; Card 3, 4, 2; etc. till all six ‘cards’. But I made a difference between search products (smartphone) and experience products (computer game). For both product categories I did exactly the same. Thus now I have two times an order of 6 ‘cards’ with two answers each. I also asked their sex, age and three introduction questions with a Likert scale of 1 to 5 before I conducted the tests with the ‘cards’. What I want to find out if there is a difference between the choices of cards that the participants made for search and experience products. Thus I want to compare them to each other. Do people make other decisions for a search product than for an experience product. However I have no idea how to do it. My supervisor told me to use Anova in SPSS. I’ve checked some videos and manuals. But I don’t know how I can test, what I want to test. I put my data in SPSS in the following manner: Variable 1 – Age Variable 2 – Sex Variable 3 – Introduction question 1 Variable 4 – Introduction question 2 Variable 5 – Introduction question 3 Variable 6 – Search product or Experience product Variable 7 – First choice (Thus did you choose card 1, 2, 3, 4, 5 or 6) Variable 8 – Credibility 1 Variable 9 – Where to buy (in a store or online) 1 Variable 10 – Second choice (Thus card 1 till 6 with exception of your first choice of course) Variable 11 – Credibility 2 Variable 12 – Where to buy 2 Variable 13 – Third choice (Thus card 1 till 6 with exception of the first two choices of course) Etc., till all six cards are gone. Thus what I want to find out is, if the order of cards people choose will be different for search and experience products, the credibility and where to buy this product for every card. I wish you all a happy new year and I hope that one of you could help me out! Thanks in advance! Freek Verzonden vanuit Mail voor Windows 10 |
Administrator
|
I'm not going to get deep at all but would simply suggest that this will be MUCH easier to manage in LONG format rather than WIDE! Since it sounds like the data are already entered I will direct you to the VARSTOCASES command in the FM (Fine Manual which nobody reads ;-) and since you are probably a point and clicker From the menu : Data > Restructure > Restructure Selected Variables into Cases ......
Be sure to have a saved copy of your data before fooling around with this. Have a great New Year.
Please reply to the list and not to my personal email.
Those desiring my consulting or training services please feel free to email me. --- "Nolite dare sanctum canibus neque mittatis margaritas vestras ante porcos ne forte conculcent eas pedibus suis." Cum es damnatorum possederunt porcos iens ut salire off sanguinum cliff in abyssum?" |
|
In reply to this post by freek kuper
Hi How about you shoot your supervisor for letting you get this far? With only 30 cases in a (non-probability?) sample, I doubt there’s much you can do beyond descriptive statistics. You don’t say whether you have ever used SPSS. If not you could start with the beginner level tutorials on my site. Other listers can advise on ANOVA, but you can do something with cross-tabulation: the problem there will be small cell sizes. John F Hall (Mr) [Retired academic survey researcher] Email: [hidden email] Website: www.surveyresearch.weebly.com SPSS start page: www.surveyresearch.weebly.com/1-survey-analysis-workshop From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of freek kuper Hi all! I’m stuck with my report. I collected some data, but now I don’t know what to do in SPSS. First some background about my report and data collection: I interviewed 30 people. They had to choose between 6 ‘cards’. Every card represent a product. This way I find out which product my participants would buy. If they choose one ‘card’, I asked two questions. One question about credibility and one about where to buy this product. Both with a Likert scale of 1 to 5. I did this until there were no ‘cards’ left. Thus the results are something like this for example: Card 5, 3, 4; Card 1, 2, 2; Card 3, 4, 2; etc. till all six ‘cards’. But I made a difference between search products (smartphone) and experience products (computer game). For both product categories I did exactly the same. Thus now I have two times an order of 6 ‘cards’ with two answers each. I also asked their sex, age and three introduction questions with a Likert scale of 1 to 5 before I conducted the tests with the ‘cards’. What I want to find out if there is a difference between the choices of cards that the participants made for search and experience products. Thus I want to compare them to each other. Do people make other decisions for a search product than for an experience product. However I have no idea how to do it. My supervisor told me to use Anova in SPSS. I’ve checked some videos and manuals. But I don’t know how I can test, what I want to test. I put my data in SPSS in the following manner: Variable 1 – Age Variable 2 – Sex Variable 3 – Introduction question 1 Variable 4 – Introduction question 2 Variable 5 – Introduction question 3 Variable 6 – Search product or Experience product Variable 7 – First choice (Thus did you choose card 1, 2, 3, 4, 5 or 6) Variable 8 – Credibility 1 Variable 9 – Where to buy (in a store or online) 1 Variable 10 – Second choice (Thus card 1 till 6 with exception of your first choice of course) Variable 11 – Credibility 2 Variable 12 – Where to buy 2 Variable 13 – Third choice (Thus card 1 till 6 with exception of the first two choices of course) Etc., till all six cards are gone. Thus what I want to find out is, if the order of cards people choose will be different for search and experience products, the credibility and where to buy this product for every card. I wish you all a happy new year and I hope that one of you could help me out! Thanks in advance! Freek Verzonden vanuit Mail voor Windows 10 ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD |
Re: Beginner is stuck with SPSS
|
|
In reply to this post by freek kuper
I think I understand the basic design: Each person picks one of six cards and answers two questions about that card. Repeat that process for the five remaining cards. So your data is a ranking of the cards plus
the two followup questions about each card. However, I’m confused by this sentence, “But I made a difference between search products (smartphone) and experience products (computer game). For both product categories I did exactly
the same. Thus now I have two times an order of 6 ‘cards’ with two answers each.” Please explain this sentence in operational/behavioral terms. Your placement of this data value (variable 6) in the data layout you included
(Thank you very much for that) suggests that your people declared “Search” or “Experience” before you showed them the set of cards. So, the value of variable 6 applies to the set of responses that follow, i.e., variables 7-24. But, then you say this, “Thus
now I have two times an order of 6 ‘cards’ with two answers each.” Maybe there’s really a different story.
Confusion in à confusion out. Gene Maguin From: SPSSX(r) Discussion [mailto:[hidden email]]
On Behalf Of freek kuper Hi all! I’m stuck with my report. I collected some data, but now I don’t know what to do in SPSS. First some background about my report and data collection: I interviewed 30 people. They had to choose between 6 ‘cards’. Every card represent a product. This way I find out which product my participants would buy. If they choose one ‘card’, I asked two questions. One question about
credibility and one about where to buy this product. Both with a Likert scale of 1 to 5. I did this until there were no ‘cards’ left. Thus the results are something like this for example: Card 5, 3, 4; Card 1, 2, 2; Card 3, 4, 2; etc. till all six ‘cards’.
But I made a difference between search products (smartphone) and experience products (computer game). For both product categories I did exactly the same. Thus now I have two times an order of 6 ‘cards’ with two answers each.
I also asked their sex, age and three introduction questions with a Likert scale of 1 to 5 before I conducted the tests with the ‘cards’.
What I want to find out if there is a difference between the choices of cards that the participants made for search and experience products. Thus I want to compare them to each other. Do people make
other decisions for a search product than for an experience product. However I have no idea how to do it. My supervisor told me to use Anova in SPSS. I’ve checked some videos and manuals. But I don’t know how I can test, what I want to test. I put my data in SPSS in the following manner:
Variable 1 – Age Variable 2 – Sex Variable 3 – Introduction question 1 Variable 4 – Introduction question 2 Variable 5 – Introduction question 3 Variable 6 – Search product or Experience product Variable 7 – First choice (Thus did you choose card 1, 2, 3, 4, 5 or 6) Variable 8 – Credibility 1 Variable 9 – Where to buy (in a store or online) 1 Variable 10 – Second choice (Thus card 1 till 6 with exception of your first choice of course) Variable 11 – Credibility 2 Variable 12 – Where to buy 2 Variable 13 – Third choice (Thus card 1 till 6 with exception of the first two choices of course) Etc., till all six cards are gone. Thus what I want to find out is, if the order of cards people choose will be different for search and experience products, the credibility and where to buy this product for every card.
I wish you all a happy new year and I hope that one of you could help me out! Thanks in advance! Freek Verzonden vanuit
Mail voor Windows 10 ===================== To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD
|
Re: Beginner is stuck with SPSS
|
|
In reply to this post by freek kuper
Your reply was very helpful. Thank you.
You expect that more people will choose card 2 first.
It seems to me that that this is simply a one sample chisquare test or a binomial test (in spss: Npar tests). Choosing one of six cards is like rolling one dice. In a fair dice the long run expectation is that
‘2’ comes up 16.6…% of the time. Same here. So, your test is card 2 first versus any other card first with an expected value of .166…. Gene Maguin From: freek kuper [mailto:[hidden email]] Hi First of all thanks for all the replies! I do have some experience with SPSS. But then I didn’t put the data in my self. I did a few simple tests, T-tests, regressions, put variables together etc., I think most of the basic stuff. I even had an exam about this. But now I have
to put the data in my self and find out what tests I should use. There is indeed a ranking of the cards plus the two follow up questions.
I’ll try to explain this sentence: “But I made a difference between search products (smartphone) and experience products (computer game). For both product categories I did exactly the same. Thus now I have two times an order of 6 ‘cards’ with two answers
each.” Every person first saw the six cards for the smartphone (thus the search product). I now have a ranking of the cards for the search product and the two follow up questions. Then I showed them the cards for the computer game (thus the experience
product). Now I have also a ranking of six cards with two follow up questions each. Thus now I have 30 responses for a search good and 30 responses (the same participants) for the experience product.
I hope this explains it. Now I want to see what the ranking is. In one of my hypotheses, based on scientific articles, I said that people will pick card 2 as the first choice. I want to check if this is significant (if that is even possible). But I might also be
able to this with just the descriptives. But then I also want to compare the rankings of the search goods with the rankings of the experience goods. To see if there is a difference. Another hypotheses of my is that people will choose for more positive reviews
for an experience good then for a search good. I hope I explained it a bit better now. Thanks again! Freek Verzonden vanuit
Mail voor Windows 10
I think I understand the basic design: Each person picks one of six cards and answers two questions about that card. Repeat that process for the five remaining cards. So your data is a ranking of the cards plus
the two followup questions about each card. However, I’m confused by this sentence, “But I made a difference between search products (smartphone) and experience products (computer game). For both product categories I did exactly
the same. Thus now I have two times an order of 6 ‘cards’ with two answers each.” Please explain this sentence in operational/behavioral terms. Your placement of this data value (variable 6) in the data layout you included
(Thank you very much for that) suggests that your people declared “Search” or “Experience” before you showed them the set of cards. So, the value of variable 6 applies to the set of responses that follow, i.e., variables 7-24. But, then you say this, “Thus
now I have two times an order of 6 ‘cards’ with two answers each.” Maybe there’s really a different story.
Confusion in à confusion out. Gene Maguin From: SPSSX(r) Discussion [[hidden email]]
On Behalf Of freek kuper Hi all! I’m stuck with my report. I collected some data, but now I don’t know what to do in SPSS. First some background about my report and data collection: I interviewed 30 people. They had to choose between 6 ‘cards’. Every card represent a product. This way I find out which product my participants would buy. If they choose one ‘card’, I asked two questions. One question about
credibility and one about where to buy this product. Both with a Likert scale of 1 to 5. I did this until there were no ‘cards’ left. Thus the results are something like this for example: Card 5, 3, 4; Card 1, 2, 2; Card 3, 4, 2; etc. till all six ‘cards’.
But I made a difference between search products (smartphone) and experience products (computer game). For both product categories I did exactly the same. Thus now I have two times an order of 6 ‘cards’ with two answers each.
I also asked their sex, age and three introduction questions with a Likert scale of 1 to 5 before I conducted the tests with the ‘cards’.
What I want to find out if there is a difference between the choices of cards that the participants made for search and experience products. Thus I want to compare them to each other. Do people make
other decisions for a search product than for an experience product. However I have no idea how to do it. My supervisor told me to use Anova in SPSS. I’ve checked some videos and manuals. But I don’t know how I can test, what I want to test. I put my data in SPSS in the following manner:
Variable 1 – Age Variable 2 – Sex Variable 3 – Introduction question 1 Variable 4 – Introduction question 2 Variable 5 – Introduction question 3 Variable 6 – Search product or Experience product Variable 7 – First choice (Thus did you choose card 1, 2, 3, 4, 5 or 6) Variable 8 – Credibility 1 Variable 9 – Where to buy (in a store or online) 1 Variable 10 – Second choice (Thus card 1 till 6 with exception of your first choice of course) Variable 11 – Credibility 2 Variable 12 – Where to buy 2 Variable 13 – Third choice (Thus card 1 till 6 with exception of the first two choices of course) Etc., till all six cards are gone. Thus what I want to find out is, if the order of cards people choose will be different for search and experience products, the credibility and where to buy this product for every card.
I wish you all a happy new year and I hope that one of you could help me out! Thanks in advance! Freek Verzonden vanuit
Mail voor Windows 10 ===================== To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text
except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD
===================== To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the command.
To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD
|
|
|
Gene One of Freek’s problems was that each of his 30 cases had two lines of data, one for search and one for game, making 60 cases instead. Also data for sex, age and online review and purchase behaviour were only in the first line. I’ve helped him to re-organise the matrix so that there are now 30 cases, with all the Smartphone data in one set of variables and all the computer game data in another set. This still leaves a logical problem when trying to analyse both sets together. That would involve some complex programming for new (temporary?) variables to contain data for search and game in the same table. Given his deadlines it could well be quicker for him to produce separate tables, copy the counts manually into a separate table and then do the stats by hand. David suggested VARSTOCASES, but I’m not sure if this would actually help. Happy New Year as well John John F Hall (Mr) [Retired academic survey researcher] Email: [hidden email] Website: www.surveyresearch.weebly.com SPSS start page: www.surveyresearch.weebly.com/1-survey-analysis-workshop From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Maguin, Eugene Your reply was very helpful. Thank you. You expect that more people will choose card 2 first. It seems to me that that this is simply a one sample chisquare test or a binomial test (in spss: Npar tests). Choosing one of six cards is like rolling one dice. In a fair dice the long run expectation is that ‘2’ comes up 16.6…% of the time. Same here. So, your test is card 2 first versus any other card first with an expected value of .166…. Gene Maguin From: freek kuper [[hidden email]] Hi First of all thanks for all the replies! I do have some experience with SPSS. But then I didn’t put the data in my self. I did a few simple tests, T-tests, regressions, put variables together etc., I think most of the basic stuff. I even had an exam about this. But now I have to put the data in my self and find out what tests I should use. There is indeed a ranking of the cards plus the two follow up questions. I’ll try to explain this sentence: “But I made a difference between search products (smartphone) and experience products (computer game). For both product categories I did exactly the same. Thus now I have two times an order of 6 ‘cards’ with two answers each.” Every person first saw the six cards for the smartphone (thus the search product). I now have a ranking of the cards for the search product and the two follow up questions. Then I showed them the cards for the computer game (thus the experience product). Now I have also a ranking of six cards with two follow up questions each. Thus now I have 30 responses for a search good and 30 responses (the same participants) for the experience product. I hope this explains it. Now I want to see what the ranking is. In one of my hypotheses, based on scientific articles, I said that people will pick card 2 as the first choice. I want to check if this is significant (if that is even possible). But I might also be able to this with just the descriptives. But then I also want to compare the rankings of the search goods with the rankings of the experience goods. To see if there is a difference. Another hypotheses of my is that people will choose for more positive reviews for an experience good then for a search good. I hope I explained it a bit better now. Thanks again! Freek Verzonden vanuit Mail voor Windows 10
I think I understand the basic design: Each person picks one of six cards and answers two questions about that card. Repeat that process for the five remaining cards. So your data is a ranking of the cards plus the two followup questions about each card. However, I’m confused by this sentence, “But I made a difference between search products (smartphone) and experience products (computer game). For both product categories I did exactly the same. Thus now I have two times an order of 6 ‘cards’ with two answers each.” Please explain this sentence in operational/behavioral terms. Your placement of this data value (variable 6) in the data layout you included (Thank you very much for that) suggests that your people declared “Search” or “Experience” before you showed them the set of cards. So, the value of variable 6 applies to the set of responses that follow, i.e., variables 7-24. But, then you say this, “Thus now I have two times an order of 6 ‘cards’ with two answers each.” Maybe there’s really a different story. Confusion in à confusion out. Gene Maguin From: SPSSX(r) Discussion [[hidden email]] On Behalf Of freek kuper Hi all! I’m stuck with my report. I collected some data, but now I don’t know what to do in SPSS. First some background about my report and data collection: I interviewed 30 people. They had to choose between 6 ‘cards’. Every card represent a product. This way I find out which product my participants would buy. If they choose one ‘card’, I asked two questions. One question about credibility and one about where to buy this product. Both with a Likert scale of 1 to 5. I did this until there were no ‘cards’ left. Thus the results are something like this for example: Card 5, 3, 4; Card 1, 2, 2; Card 3, 4, 2; etc. till all six ‘cards’. But I made a difference between search products (smartphone) and experience products (computer game). For both product categories I did exactly the same. Thus now I have two times an order of 6 ‘cards’ with two answers each. I also asked their sex, age and three introduction questions with a Likert scale of 1 to 5 before I conducted the tests with the ‘cards’. What I want to find out if there is a difference between the choices of cards that the participants made for search and experience products. Thus I want to compare them to each other. Do people make other decisions for a search product than for an experience product. However I have no idea how to do it. My supervisor told me to use Anova in SPSS. I’ve checked some videos and manuals. But I don’t know how I can test, what I want to test. I put my data in SPSS in the following manner: Variable 1 – Age Variable 2 – Sex Variable 3 – Introduction question 1 Variable 4 – Introduction question 2 Variable 5 – Introduction question 3 Variable 6 – Search product or Experience product Variable 7 – First choice (Thus did you choose card 1, 2, 3, 4, 5 or 6) Variable 8 – Credibility 1 Variable 9 – Where to buy (in a store or online) 1 Variable 10 – Second choice (Thus card 1 till 6 with exception of your first choice of course) Variable 11 – Credibility 2 Variable 12 – Where to buy 2 Variable 13 – Third choice (Thus card 1 till 6 with exception of the first two choices of course) Etc., till all six cards are gone. Thus what I want to find out is, if the order of cards people choose will be different for search and experience products, the credibility and where to buy this product for every card. I wish you all a happy new year and I hope that one of you could help me out! Thanks in advance! Freek Verzonden vanuit Mail voor Windows 10 ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD |
Re: Beginner is stuck with SPSS
|
|
Yes, that fact wasn’t apparent to me in the original post. Ok, but you’ve helped him restructure is data into one record per person. I remember seeing David’s post last week but I didn’t understand the intention.
Now I do. In his/her off-list reply to me, she names a specific analysis question: card 2 more likely to be picked first. I also see that I didn’t read whole message. I missed this “But then I also want to compare the rankings of the search goods with
the rankings of the experience goods. To see if there is a difference. Another hypotheses of my is that people will choose for more positive reviews for an experience good then for a search good.” As I understand these two statements, I don’t think they are analyzable because I do not see how to relate “rankings of the search goods” to a specific variable in the dataset. Same issue for “more positive reviews”.
There’s a comparison between search goods and experience goods to be made, so a repeated measures design but more information is needed to move forward.
Gene Maguin From: SPSSX(r) Discussion [mailto:[hidden email]]
On Behalf Of John F Hall Gene One of Freek’s problems was that each of his 30 cases had two lines of data, one for search and one for game, making 60 cases instead. Also data for sex, age and online review and purchase behaviour
were only in the first line. I’ve helped him to re-organise the matrix so that there are now 30 cases, with all the Smartphone data in one set of variables and all the computer game data in another set.
This still leaves a logical problem when trying to analyse both sets together. That would involve some complex programming for new (temporary?) variables to contain data for search and game in the
same table. Given his deadlines it could well be quicker for him to produce separate tables, copy the counts manually into a separate table and then do the stats by hand. David suggested VARSTOCASES, but I’m not sure if this would actually help. Happy New Year as well John John F Hall (Mr) [Retired academic survey researcher] Email:
[hidden email]
Website:
www.surveyresearch.weebly.com
SPSS start page:
www.surveyresearch.weebly.com/1-survey-analysis-workshop From: SPSSX(r) Discussion [[hidden email]]
On Behalf Of Maguin, Eugene Your reply was very helpful. Thank you.
You expect that more people will choose card 2 first.
It seems to me that that this is simply a one sample chisquare test or a binomial test (in spss: Npar tests). Choosing one of six cards is like rolling one dice. In a fair dice the long run expectation is that
‘2’ comes up 16.6…% of the time. Same here. So, your test is card 2 first versus any other card first with an expected value of .166…. Gene Maguin From: freek kuper [[hidden email]]
Hi First of all thanks for all the replies! I do have some experience with SPSS. But then I didn’t put the data in my self. I did a few simple tests, T-tests, regressions, put variables together etc., I think most of the basic stuff. I even had an exam about this. But now I have
to put the data in my self and find out what tests I should use. There is indeed a ranking of the cards plus the two follow up questions.
I’ll try to explain this sentence: “But I made a difference between search products (smartphone) and experience products (computer game). For both product categories I did exactly the same. Thus now I have two times an order of 6 ‘cards’ with two answers
each.” Every person first saw the six cards for the smartphone (thus the search product). I now have a ranking of the cards for the search product and the two follow up questions. Then I showed them the cards for the computer game (thus the experience
product). Now I have also a ranking of six cards with two follow up questions each. Thus now I have 30 responses for a search good and 30 responses (the same participants) for the experience product.
I hope this explains it. Now I want to see what the ranking is. In one of my hypotheses, based on scientific articles, I said that people will pick card 2 as the first choice. I want to check if this is significant (if that is even possible). But I might also be
able to this with just the descriptives. But then I also want to compare the rankings of the search goods with the rankings of the experience goods. To see if there is a difference. Another hypotheses of my is that people will choose for more positive reviews
for an experience good then for a search good. I hope I explained it a bit better now. Thanks again! Freek Verzonden vanuit
Mail voor Windows 10
I think I understand the basic design: Each person picks one of six cards and answers two questions about that card. Repeat that process for the five remaining cards. So your data is a ranking of the cards plus
the two followup questions about each card. However, I’m confused by this sentence, “But I made a difference between search products (smartphone) and experience products (computer game). For both product categories I did exactly
the same. Thus now I have two times an order of 6 ‘cards’ with two answers each.” Please explain this sentence in operational/behavioral terms. Your placement of this data value (variable 6) in the data layout you included
(Thank you very much for that) suggests that your people declared “Search” or “Experience” before you showed them the set of cards. So, the value of variable 6 applies to the set of responses that follow, i.e., variables 7-24. But, then you say this, “Thus
now I have two times an order of 6 ‘cards’ with two answers each.” Maybe there’s really a different story.
Confusion in à confusion out. Gene Maguin From: SPSSX(r) Discussion [[hidden email]]
On Behalf Of freek kuper Hi all! I’m stuck with my report. I collected some data, but now I don’t know what to do in SPSS. First some background about my report and data collection: I interviewed 30 people. They had to choose between 6 ‘cards’. Every card represent a product. This way I find out which product my participants would buy. If they choose one ‘card’, I asked two questions. One question about
credibility and one about where to buy this product. Both with a Likert scale of 1 to 5. I did this until there were no ‘cards’ left. Thus the results are something like this for example: Card 5, 3, 4; Card 1, 2, 2; Card 3, 4, 2; etc. till all six ‘cards’.
But I made a difference between search products (smartphone) and experience products (computer game). For both product categories I did exactly the same. Thus now I have two times an order of 6 ‘cards’ with two answers each.
I also asked their sex, age and three introduction questions with a Likert scale of 1 to 5 before I conducted the tests with the ‘cards’.
What I want to find out if there is a difference between the choices of cards that the participants made for search and experience products. Thus I want to compare them to each other. Do people make
other decisions for a search product than for an experience product. However I have no idea how to do it. My supervisor told me to use Anova in SPSS. I’ve checked some videos and manuals. But I don’t know how I can test, what I want to test. I put my data in SPSS in the following manner:
Variable 1 – Age Variable 2 – Sex Variable 3 – Introduction question 1 Variable 4 – Introduction question 2 Variable 5 – Introduction question 3 Variable 6 – Search product or Experience product Variable 7 – First choice (Thus did you choose card 1, 2, 3, 4, 5 or 6) Variable 8 – Credibility 1 Variable 9 – Where to buy (in a store or online) 1 Variable 10 – Second choice (Thus card 1 till 6 with exception of your first choice of course) Variable 11 – Credibility 2 Variable 12 – Where to buy 2 Variable 13 – Third choice (Thus card 1 till 6 with exception of the first two choices of course) Etc., till all six cards are gone. Thus what I want to find out is, if the order of cards people choose will be different for search and experience products, the credibility and where to buy this product for every card.
I wish you all a happy new year and I hope that one of you could help me out! Thanks in advance! Freek Verzonden vanuit
Mail voor Windows 10 ===================== To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text
except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD
===================== To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the command.
To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD
===================== To manage your subscription to SPSSX-L, send a message to
[hidden email]
(not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD
===================== To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text
except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD |
Re: Beginner is stuck with SPSS
|
|
In reply to this post by freek kuper
Freek, Please send replies to the list rather than to me personally. There very well might be somebody on the list who has experience with this sort of topic, experimental design structure, and analysis. They’ll have
a lot better understanding than I do. I don’t really understand the story that this topic/experimental design is getting at. That said, where is the variable
eWOM in your table? For a given respondent, which of these statements is correct about eWOM? A) It could have a different value for every card selection trial within search and within experience? B) It has the same value for every card selection trial
within search and the same value for every card selection trial within experience? C) It has the same value for all card selection trials. Gene Maguin From: freek kuper [mailto:[hidden email]] Hi! I’ve figured it out! Most of it. It’s probably very easy for you guys after this, however I’m still struggling.
I’ve reorganized my data again. I’ll set up an example for one participant below:
This times 30 (30 participants). Explanation: This way it is possible to day a Linear Regression. With Dependent variables: priority, credibility and off/online. And with independent variables: Volume and %neg. Q1, Q2 and Q3 are my control variables. ID: the participant. Age: Age of the participant. Just to explain my sample a bit more in my report. Same for gender. Q1: The question whether the participant always reads online consumer reviews (eWOM) before purchasing a product offline. (Likert scale 1 to 5) Q2: If eWOM is positive than I buy the product offline (Likert scale 1 to 5) Q3: eWOM has a crucial effect on my offline buying behavior (Likert scale 1 to 5) SvsE whether we did the experiment for a search product (phone) or an experience product (computer game) Priority: First choice till last choice. Thus first choice, second choice…. Sixth choice. Card: Which card he/she picked. Is just there to make the data more clear for me. Credibility: How credible does the participant find the chosen card (Likert scale 1 to 5) Off/online: Does the participant buy this product online (1) or offline (5) (Likert scale 1 to 5) Volume: On the cards I had three cards with a volume of 5 reviews and 3 cards with a volume of 10 reviews. %neg: Is the percentage of the negative reviews on the cards. This could be 0%, 20% and 40%. My table for the cards looked like this:
The plan is to do 3*3 times a linear regression. For the dependent variables each (priority, credibility and off/online) and then also for search and experience together, only search and only experience goods (with the help of the select
cases option). Oh yeah. It is possible to do a linear regression, because the dependent variables, which are actually ordinal, can be categorized as interval data, because they use a Likert scale.
But now I did not include my control variables? Where and how do I include those? But that is something that I probably can find easily on the internet or in my new book (which is pretty good if you ask me). Title: Discovering Statistics
Using IBM SPSS Statistics by Andy Field. 4th Edition. (I don’t know if any of you know/use this book and what you guys think of it, but I can recommend it!) However tips are always welcome! If I did anything wrong, would you please notice me! Thanks a lot for all your help guys! Especially you John! Many thanks for all the time you have invested. I hope I can do something back for you someday! Freek Verzonden vanuit
Mail voor Windows 10
Yes, that fact wasn’t apparent to me in the original post. Ok, but you’ve helped him restructure is data into one record per person. I remember seeing David’s post last week but I didn’t understand the intention.
Now I do. In his/her off-list reply to me, she names a specific analysis question: card 2 more likely to be picked first. I also see that I didn’t read whole message. I missed this “But then I also want to compare the rankings of the search goods with
the rankings of the experience goods. To see if there is a difference. Another hypotheses of my is that people will choose for more positive reviews for an experience good then for a search good.” As I understand these two statements, I don’t think they are analyzable because I do not see how to relate “rankings of the search goods” to a specific variable in the dataset. Same issue for “more positive reviews”.
There’s a comparison between search goods and experience goods to be made, so a repeated measures design but more information is needed to move forward.
Gene Maguin From: SPSSX(r) Discussion [[hidden email]]
On Behalf Of John F Hall Gene One of Freek’s problems was that each of his 30 cases had two lines of data, one for search and one for game, making 60 cases instead. Also data for sex, age and online review and purchase behaviour
were only in the first line. I’ve helped him to re-organise the matrix so that there are now 30 cases, with all the Smartphone data in one set of variables and all the computer game data in another set.
This still leaves a logical problem when trying to analyse both sets together. That would involve some complex programming for new (temporary?) variables to contain data for search and game in the
same table. Given his deadlines it could well be quicker for him to produce separate tables, copy the counts manually into a separate table and then do the stats by hand. David suggested VARSTOCASES, but I’m not sure if this would actually help. Happy New Year as well John John F Hall (Mr) [Retired academic survey researcher] Email:
[hidden email]
Website:
www.surveyresearch.weebly.com
SPSS start page:
www.surveyresearch.weebly.com/1-survey-analysis-workshop From: SPSSX(r) Discussion [[hidden email]]
On Behalf Of Maguin, Eugene Your reply was very helpful. Thank you.
You expect that more people will choose card 2 first.
It seems to me that that this is simply a one sample chisquare test or a binomial test (in spss: Npar tests). Choosing one of six cards is like rolling one dice. In a fair dice the long run expectation is that
‘2’ comes up 16.6…% of the time. Same here. So, your test is card 2 first versus any other card first with an expected value of .166…. Gene Maguin From: freek kuper [[hidden email]]
Hi First of all thanks for all the replies! I do have some experience with SPSS. But then I didn’t put the data in my self. I did a few simple tests, T-tests, regressions, put variables together etc., I think most of the basic stuff. I even had an exam about this. But now I have
to put the data in my self and find out what tests I should use. There is indeed a ranking of the cards plus the two follow up questions.
I’ll try to explain this sentence: “But I made a difference between search products (smartphone) and experience products (computer game). For both product categories I did exactly the same. Thus now I have two times an order of 6 ‘cards’ with two answers
each.” Every person first saw the six cards for the smartphone (thus the search product). I now have a ranking of the cards for the search product and the two follow up questions. Then I showed them the cards for the computer game (thus the experience
product). Now I have also a ranking of six cards with two follow up questions each. Thus now I have 30 responses for a search good and 30 responses (the same participants) for the experience product.
I hope this explains it. Now I want to see what the ranking is. In one of my hypotheses, based on scientific articles, I said that people will pick card 2 as the first choice. I want to check if this is significant (if that is even possible). But I might also be
able to this with just the descriptives. But then I also want to compare the rankings of the search goods with the rankings of the experience goods. To see if there is a difference. Another hypotheses of my is that people will choose for more positive reviews
for an experience good then for a search good. I hope I explained it a bit better now. Thanks again! Freek Verzonden vanuit
Mail voor Windows 10
I think I understand the basic design: Each person picks one of six cards and answers two questions about that card. Repeat that process for the five remaining cards. So your data is a ranking of the cards plus
the two followup questions about each card. However, I’m confused by this sentence, “But I made a difference between search products (smartphone) and experience products (computer game). For both product categories I did exactly
the same. Thus now I have two times an order of 6 ‘cards’ with two answers each.” Please explain this sentence in operational/behavioral terms. Your placement of this data value (variable 6) in the data layout you included
(Thank you very much for that) suggests that your people declared “Search” or “Experience” before you showed them the set of cards. So, the value of variable 6 applies to the set of responses that follow, i.e., variables 7-24. But, then you say this, “Thus
now I have two times an order of 6 ‘cards’ with two answers each.” Maybe there’s really a different story.
Confusion in à confusion out. Gene Maguin From: SPSSX(r) Discussion [[hidden email]]
On Behalf Of freek kuper Hi all! I’m stuck with my report. I collected some data, but now I don’t know what to do in SPSS. First some background about my report and data collection: I interviewed 30 people. They had to choose between 6 ‘cards’. Every card represent a product. This way I find out which product my participants would buy. If they choose one ‘card’, I asked two questions. One question about
credibility and one about where to buy this product. Both with a Likert scale of 1 to 5. I did this until there were no ‘cards’ left. Thus the results are something like this for example: Card 5, 3, 4; Card 1, 2, 2; Card 3, 4, 2; etc. till all six ‘cards’.
But I made a difference between search products (smartphone) and experience products (computer game). For both product categories I did exactly the same. Thus now I have two times an order of 6 ‘cards’ with two answers each.
I also asked their sex, age and three introduction questions with a Likert scale of 1 to 5 before I conducted the tests with the ‘cards’.
What I want to find out if there is a difference between the choices of cards that the participants made for search and experience products. Thus I want to compare them to each other. Do people make
other decisions for a search product than for an experience product. However I have no idea how to do it. My supervisor told me to use Anova in SPSS. I’ve checked some videos and manuals. But I don’t know how I can test, what I want to test. I put my data in SPSS in the following manner:
Variable 1 – Age Variable 2 – Sex Variable 3 – Introduction question 1 Variable 4 – Introduction question 2 Variable 5 – Introduction question 3 Variable 6 – Search product or Experience product Variable 7 – First choice (Thus did you choose card 1, 2, 3, 4, 5 or 6) Variable 8 – Credibility 1 Variable 9 – Where to buy (in a store or online) 1 Variable 10 – Second choice (Thus card 1 till 6 with exception of your first choice of course) Variable 11 – Credibility 2 Variable 12 – Where to buy 2 Variable 13 – Third choice (Thus card 1 till 6 with exception of the first two choices of course) Etc., till all six cards are gone. Thus what I want to find out is, if the order of cards people choose will be different for search and experience products, the credibility and where to buy this product for every card.
I wish you all a happy new year and I hope that one of you could help me out! Thanks in advance! Freek Verzonden vanuit
Mail voor Windows 10 ===================== To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text
except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD
===================== To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the command.
To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD
===================== To manage your subscription to SPSSX-L, send a message to
[hidden email]
(not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command
INFO REFCARD ===================== To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text
except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD ===================== To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD
|
|
|
In reply to this post by freek kuper
Hi, Eugene I thought I mailed it to the list. To answer your question. eWOM is in the table as ‘volume’ and ‘%negative’. I try to figure out what these effects have on consumer buying behavior in physical stores. It doesn’t matter what there is told in the review. It’s about negative/positive and volume in their first look. Thus does a consumer see 20 positive reviews and 3 positive reviews for example, which product does the consumer buy. So it’s not about the content of the eWOM, but about the layout and if it’s positive or negative. That’s why I marked the positive reviews black and negative reviews red on my cards. Here is the mail I accidently only send to Eugene (but I think he put it in the list by answering me, not sure though): I’ve figured it out! Most of it. It’s probably very easy for you guys after this, however I’m still struggling. I’ve reorganized my data again. I’ll set up an example for one participant below:
This times 30 (30 participants). Explanation: This way it is possible to day a Linear Regression. With Dependent variables: priority, credibility and off/online. And with independent variables: Volume and %neg. Q1, Q2 and Q3 are my control variables. ID: the participant. Age: Age of the participant. Just to explain my sample a bit more in my report. Same for gender. Q1: The question whether the participant always reads online consumer reviews (eWOM) before purchasing a product offline. (Likert scale 1 to 5) Q2: If eWOM is positive than I buy the product offline (Likert scale 1 to 5) Q3: eWOM has a crucial effect on my offline buying behavior (Likert scale 1 to 5) SvsE whether we did the experiment for a search product (phone) or an experience product (computer game) Priority: First choice till last choice. Thus first choice, second choice…. Sixth choice. Card: Which card he/she picked. Is just there to make the data more clear for me. Credibility: How credible does the participant find the chosen card (Likert scale 1 to 5) Off/online: Does the participant buy this product online (1) or offline (5) (Likert scale 1 to 5) Volume: On the cards I had three cards with a volume of 5 reviews and 3 cards with a volume of 10 reviews. %neg: Is the percentage of the negative reviews on the cards. This could be 0%, 20% and 40%. My table for the cards looked like this:
The plan is to do 3*3 times a linear regression. For the dependent variables each (priority, credibility and off/online) and then also for search and experience together, only search and only experience goods (with the help of the select cases option). Oh yeah. It is possible to do a linear regression, because the dependent variables, which are actually ordinal, can be categorized as interval data, because they use a Likert scale. But now I did not include my control variables? Where and how do I include those? But that is something that I probably can find easily on the internet or in my new book (which is pretty good if you ask me). Title: Discovering Statistics Using IBM SPSS Statistics by Andy Field. 4th Edition. (I don’t know if any of you know/use this book and what you guys think of it, but I can recommend it!) However tips are always welcome! If I did anything wrong, would you please notice me! Thanks a lot for all your help guys! Especially you John! Many thanks for all the time you have invested. I hope I can do something back for you someday! Freek Verzonden vanuit Mail voor Windows 10
Freek, Please send replies to the list rather than to me personally. There very well might be somebody on the list who has experience with this sort of topic, experimental design structure, and analysis. They’ll have a lot better understanding than I do. I don’t really understand the story that this topic/experimental design is getting at. That said, where is the variable eWOM in your table? For a given respondent, which of these statements is correct about eWOM? A) It could have a different value for every card selection trial within search and within experience? B) It has the same value for every card selection trial within search and the same value for every card selection trial within experience? C) It has the same value for all card selection trials. Gene Maguin From: freek kuper [mailto:[hidden email]] Hi! I’ve figured it out! Most of it. It’s probably very easy for you guys after this, however I’m still struggling. I’ve reorganized my data again. I’ll set up an example for one participant below:
This times 30 (30 participants). Explanation: This way it is possible to day a Linear Regression. With Dependent variables: priority, credibility and off/online. And with independent variables: Volume and %neg. Q1, Q2 and Q3 are my control variables. ID: the participant. Age: Age of the participant. Just to explain my sample a bit more in my report. Same for gender. Q1: The question whether the participant always reads online consumer reviews (eWOM) before purchasing a product offline. (Likert scale 1 to 5) Q2: If eWOM is positive than I buy the product offline (Likert scale 1 to 5) Q3: eWOM has a crucial effect on my offline buying behavior (Likert scale 1 to 5) SvsE whether we did the experiment for a search product (phone) or an experience product (computer game) Priority: First choice till last choice. Thus first choice, second choice…. Sixth choice. Card: Which card he/she picked. Is just there to make the data more clear for me. Credibility: How credible does the participant find the chosen card (Likert scale 1 to 5) Off/online: Does the participant buy this product online (1) or offline (5) (Likert scale 1 to 5) Volume: On the cards I had three cards with a volume of 5 reviews and 3 cards with a volume of 10 reviews. %neg: Is the percentage of the negative reviews on the cards. This could be 0%, 20% and 40%. My table for the cards looked like this:
The plan is to do 3*3 times a linear regression. For the dependent variables each (priority, credibility and off/online) and then also for search and experience together, only search and only experience goods (with the help of the select cases option). Oh yeah. It is possible to do a linear regression, because the dependent variables, which are actually ordinal, can be categorized as interval data, because they use a Likert scale. But now I did not include my control variables? Where and how do I include those? But that is something that I probably can find easily on the internet or in my new book (which is pretty good if you ask me). Title: Discovering Statistics Using IBM SPSS Statistics by Andy Field. 4th Edition. (I don’t know if any of you know/use this book and what you guys think of it, but I can recommend it!) However tips are always welcome! If I did anything wrong, would you please notice me! Thanks a lot for all your help guys! Especially you John! Many thanks for all the time you have invested. I hope I can do something back for you someday! Freek Verzonden vanuit Mail voor Windows 10
Yes, that fact wasn’t apparent to me in the original post. Ok, but you’ve helped him restructure is data into one record per person. I remember seeing David’s post last week but I didn’t understand the intention. Now I do. In his/her off-list reply to me, she names a specific analysis question: card 2 more likely to be picked first. I also see that I didn’t read whole message. I missed this “But then I also want to compare the rankings of the search goods with the rankings of the experience goods. To see if there is a difference. Another hypotheses of my is that people will choose for more positive reviews for an experience good then for a search good.” As I understand these two statements, I don’t think they are analyzable because I do not see how to relate “rankings of the search goods” to a specific variable in the dataset. Same issue for “more positive reviews”. There’s a comparison between search goods and experience goods to be made, so a repeated measures design but more information is needed to move forward. Gene Maguin From: SPSSX(r) Discussion [[hidden email]] On Behalf Of John F Hall Gene One of Freek’s problems was that each of his 30 cases had two lines of data, one for search and one for game, making 60 cases instead. Also data for sex, age and online review and purchase behaviour were only in the first line. I’ve helped him to re-organise the matrix so that there are now 30 cases, with all the Smartphone data in one set of variables and all the computer game data in another set. This still leaves a logical problem when trying to analyse both sets together. That would involve some complex programming for new (temporary?) variables to contain data for search and game in the same table. Given his deadlines it could well be quicker for him to produce separate tables, copy the counts manually into a separate table and then do the stats by hand. David suggested VARSTOCASES, but I’m not sure if this would actually help. Happy New Year as well John John F Hall (Mr) [Retired academic survey researcher] Email: [hidden email] Website: www.surveyresearch.weebly.com SPSS start page: www.surveyresearch.weebly.com/1-survey-analysis-workshop From: SPSSX(r) Discussion [[hidden email]] On Behalf Of Maguin, Eugene Your reply was very helpful. Thank you. You expect that more people will choose card 2 first. It seems to me that that this is simply a one sample chisquare test or a binomial test (in spss: Npar tests). Choosing one of six cards is like rolling one dice. In a fair dice the long run expectation is that ‘2’ comes up 16.6…% of the time. Same here. So, your test is card 2 first versus any other card first with an expected value of .166…. Gene Maguin From: freek kuper [[hidden email]] Hi First of all thanks for all the replies! I do have some experience with SPSS. But then I didn’t put the data in my self. I did a few simple tests, T-tests, regressions, put variables together etc., I think most of the basic stuff. I even had an exam about this. But now I have to put the data in my self and find out what tests I should use. There is indeed a ranking of the cards plus the two follow up questions. I’ll try to explain this sentence: “But I made a difference between search products (smartphone) and experience products (computer game). For both product categories I did exactly the same. Thus now I have two times an order of 6 ‘cards’ with two answers each.” Every person first saw the six cards for the smartphone (thus the search product). I now have a ranking of the cards for the search product and the two follow up questions. Then I showed them the cards for the computer game (thus the experience product). Now I have also a ranking of six cards with two follow up questions each. Thus now I have 30 responses for a search good and 30 responses (the same participants) for the experience product. I hope this explains it. Now I want to see what the ranking is. In one of my hypotheses, based on scientific articles, I said that people will pick card 2 as the first choice. I want to check if this is significant (if that is even possible). But I might also be able to this with just the descriptives. But then I also want to compare the rankings of the search goods with the rankings of the experience goods. To see if there is a difference. Another hypotheses of my is that people will choose for more positive reviews for an experience good then for a search good. I hope I explained it a bit better now. Thanks again! Freek Verzonden vanuit Mail voor Windows 10
I think I understand the basic design: Each person picks one of six cards and answers two questions about that card. Repeat that process for the five remaining cards. So your data is a ranking of the cards plus the two followup questions about each card. However, I’m confused by this sentence, “But I made a difference between search products (smartphone) and experience products (computer game). For both product categories I did exactly the same. Thus now I have two times an order of 6 ‘cards’ with two answers each.” Please explain this sentence in operational/behavioral terms. Your placement of this data value (variable 6) in the data layout you included (Thank you very much for that) suggests that your people declared “Search” or “Experience” before you showed them the set of cards. So, the value of variable 6 applies to the set of responses that follow, i.e., variables 7-24. But, then you say this, “Thus now I have two times an order of 6 ‘cards’ with two answers each.” Maybe there’s really a different story. Confusion in à confusion out. Gene Maguin From: SPSSX(r) Discussion [[hidden email]] On Behalf Of freek kuper Hi all! I’m stuck with my report. I collected some data, but now I don’t know what to do in SPSS. First some background about my report and data collection: I interviewed 30 people. They had to choose between 6 ‘cards’. Every card represent a product. This way I find out which product my participants would buy. If they choose one ‘card’, I asked two questions. One question about credibility and one about where to buy this product. Both with a Likert scale of 1 to 5. I did this until there were no ‘cards’ left. Thus the results are something like this for example: Card 5, 3, 4; Card 1, 2, 2; Card 3, 4, 2; etc. till all six ‘cards’. But I made a difference between search products (smartphone) and experience products (computer game). For both product categories I did exactly the same. Thus now I have two times an order of 6 ‘cards’ with two answers each. I also asked their sex, age and three introduction questions with a Likert scale of 1 to 5 before I conducted the tests with the ‘cards’. What I want to find out if there is a difference between the choices of cards that the participants made for search and experience products. Thus I want to compare them to each other. Do people make other decisions for a search product than for an experience product. However I have no idea how to do it. My supervisor told me to use Anova in SPSS. I’ve checked some videos and manuals. But I don’t know how I can test, what I want to test. I put my data in SPSS in the following manner: Variable 1 – Age Variable 2 – Sex Variable 3 – Introduction question 1 Variable 4 – Introduction question 2 Variable 5 – Introduction question 3 Variable 6 – Search product or Experience product Variable 7 – First choice (Thus did you choose card 1, 2, 3, 4, 5 or 6) Variable 8 – Credibility 1 Variable 9 – Where to buy (in a store or online) 1 Variable 10 – Second choice (Thus card 1 till 6 with exception of your first choice of course) Variable 11 – Credibility 2 Variable 12 – Where to buy 2 Variable 13 – Third choice (Thus card 1 till 6 with exception of the first two choices of course) Etc., till all six cards are gone. Thus what I want to find out is, if the order of cards people choose will be different for search and experience products, the credibility and where to buy this product for every card. I wish you all a happy new year and I hope that one of you could help me out! Thanks in advance! Freek Verzonden vanuit Mail voor Windows 10 ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD |
Re: Beginner is stuck with SPSS
|
|
Freek, I’ve pretty much spent as much time on this as I want to give. I still don’t understand the design narrative. Partly it’s language issues, I think, but it also seems to me that the story is shifting as either
new information is added or something changes. So, please post a citation to a published article that has used the experimental design you have used to investigate similar questions.
Gene Maguin From: SPSSX(r) Discussion [mailto:[hidden email]]
On Behalf Of freek kuper Hi, Eugene I thought I mailed it to the list. To answer your question. eWOM is in the table as ‘volume’ and ‘%negative’. I try to figure out what these effects have on consumer buying behavior in physical stores. It doesn’t matter what there is told in the review. It’s about negative/positive
and volume in their first look. Thus does a consumer see 20 positive reviews and 3 positive reviews for example, which product does the consumer buy. So it’s not about the content of the eWOM, but about the layout and if it’s positive or negative. That’s why
I marked the positive reviews black and negative reviews red on my cards. Here is the mail I accidently only send to Eugene (but I think he put it in the list by answering me, not sure though):
I’ve figured it out! Most of it. It’s probably very easy for you guys after this, however I’m still struggling.
I’ve reorganized my data again. I’ll set up an example for one participant below:
This times 30 (30 participants). Explanation: This way it is possible to day a Linear Regression. With Dependent variables: priority, credibility and off/online. And with independent variables: Volume and %neg. Q1, Q2 and Q3 are my control variables. ID: the participant. Age: Age of the participant. Just to explain my sample a bit more in my report. Same for gender. Q1: The question whether the participant always reads online consumer reviews (eWOM) before purchasing a product offline. (Likert scale 1 to 5) Q2: If eWOM is positive than I buy the product offline (Likert scale 1 to 5) Q3: eWOM has a crucial effect on my offline buying behavior (Likert scale 1 to 5) SvsE whether we did the experiment for a search product (phone) or an experience product (computer game) Priority: First choice till last choice. Thus first choice, second choice…. Sixth choice. Card: Which card he/she picked. Is just there to make the data more clear for me. Credibility: How credible does the participant find the chosen card (Likert scale 1 to 5) Off/online: Does the participant buy this product online (1) or offline (5) (Likert scale 1 to 5) Volume: On the cards I had three cards with a volume of 5 reviews and 3 cards with a volume of 10 reviews. %neg: Is the percentage of the negative reviews on the cards. This could be 0%, 20% and 40%. My table for the cards looked like this:
The plan is to do 3*3 times a linear regression. For the dependent variables each (priority, credibility and off/online) and then also for search and experience together, only search and only experience goods (with the help of the select
cases option). Oh yeah. It is possible to do a linear regression, because the dependent variables, which are actually ordinal, can be categorized as interval data, because they use a Likert scale.
But now I did not include my control variables? Where and how do I include those? But that is something that I probably can find easily on the internet or in my new book (which is pretty good if you ask me). Title: Discovering Statistics
Using IBM SPSS Statistics by Andy Field. 4th Edition. (I don’t know if any of you know/use this book and what you guys think of it, but I can recommend it!) However tips are always welcome! If I did anything wrong, would you please notice me! Thanks a lot for all your help guys! Especially you John! Many thanks for all the time you have invested. I hope I can do something back for you someday! Freek Verzonden vanuit
Mail voor Windows 10
Freek, Please send replies to the list rather than to me personally. There very well might be somebody on the list who has experience with this sort of topic, experimental design structure, and analysis. They’ll have
a lot better understanding than I do. I don’t really understand the story that this topic/experimental design is getting at. That said, where is the variable
eWOM in your table? For a given respondent, which of these statements is correct about eWOM? A) It could have a different value for every card selection trial within search and within experience? B) It has the same value for every card selection trial
within search and the same value for every card selection trial within experience? C) It has the same value for all card selection trials. Gene Maguin From: freek kuper [[hidden email]]
Hi! I’ve figured it out! Most of it. It’s probably very easy for you guys after this, however I’m still struggling.
I’ve reorganized my data again. I’ll set up an example for one participant below:
This times 30 (30 participants). Explanation: This way it is possible to day a Linear Regression. With Dependent variables: priority, credibility and off/online. And with independent variables: Volume and %neg. Q1, Q2 and Q3 are my control variables. ID: the participant. Age: Age of the participant. Just to explain my sample a bit more in my report. Same for gender. Q1: The question whether the participant always reads online consumer reviews (eWOM) before purchasing a product offline. (Likert scale 1 to 5) Q2: If eWOM is positive than I buy the product offline (Likert scale 1 to 5) Q3: eWOM has a crucial effect on my offline buying behavior (Likert scale 1 to 5) SvsE whether we did the experiment for a search product (phone) or an experience product (computer game) Priority: First choice till last choice. Thus first choice, second choice…. Sixth choice. Card: Which card he/she picked. Is just there to make the data more clear for me. Credibility: How credible does the participant find the chosen card (Likert scale 1 to 5) Off/online: Does the participant buy this product online (1) or offline (5) (Likert scale 1 to 5) Volume: On the cards I had three cards with a volume of 5 reviews and 3 cards with a volume of 10 reviews. %neg: Is the percentage of the negative reviews on the cards. This could be 0%, 20% and 40%. My table for the cards looked like this:
The plan is to do 3*3 times a linear regression. For the dependent variables each (priority, credibility and off/online) and then also for search and experience together, only search and only experience goods (with the help of the select
cases option). Oh yeah. It is possible to do a linear regression, because the dependent variables, which are actually ordinal, can be categorized as interval data, because they use a Likert scale.
But now I did not include my control variables? Where and how do I include those? But that is something that I probably can find easily on the internet or in my new book (which is pretty good if you ask me). Title: Discovering Statistics
Using IBM SPSS Statistics by Andy Field. 4th Edition. (I don’t know if any of you know/use this book and what you guys think of it, but I can recommend it!) However tips are always welcome! If I did anything wrong, would you please notice me! Thanks a lot for all your help guys! Especially you John! Many thanks for all the time you have invested. I hope I can do something back for you someday! Freek Verzonden vanuit
Mail voor Windows 10
Yes, that fact wasn’t apparent to me in the original post. Ok, but you’ve helped him restructure is data into one record per person. I remember seeing David’s post last week but I didn’t understand the intention.
Now I do. In his/her off-list reply to me, she names a specific analysis question: card 2 more likely to be picked first. I also see that I didn’t read whole message. I missed this “But then I also want to compare the rankings of the search goods with
the rankings of the experience goods. To see if there is a difference. Another hypotheses of my is that people will choose for more positive reviews for an experience good then for a search good.” As I understand these two statements, I don’t think they are analyzable because I do not see how to relate “rankings of the search goods” to a specific variable in the dataset. Same issue for “more positive reviews”.
There’s a comparison between search goods and experience goods to be made, so a repeated measures design but more information is needed to move forward.
Gene Maguin From: SPSSX(r) Discussion [[hidden email]]
On Behalf Of John F Hall Gene One of Freek’s problems was that each of his 30 cases had two lines of data, one for search and one for game, making 60 cases instead. Also data for sex, age and online review and purchase behaviour
were only in the first line. I’ve helped him to re-organise the matrix so that there are now 30 cases, with all the Smartphone data in one set of variables and all the computer game data in another set.
This still leaves a logical problem when trying to analyse both sets together. That would involve some complex programming for new (temporary?) variables to contain data for search and game in the
same table. Given his deadlines it could well be quicker for him to produce separate tables, copy the counts manually into a separate table and then do the stats by hand. David suggested VARSTOCASES, but I’m not sure if this would actually help. Happy New Year as well John John F Hall (Mr) [Retired academic survey researcher] Email:
[hidden email]
Website:
www.surveyresearch.weebly.com
SPSS start page:
www.surveyresearch.weebly.com/1-survey-analysis-workshop From: SPSSX(r) Discussion [[hidden email]]
On Behalf Of Maguin, Eugene Your reply was very helpful. Thank you.
You expect that more people will choose card 2 first.
It seems to me that that this is simply a one sample chisquare test or a binomial test (in spss: Npar tests). Choosing one of six cards is like rolling one dice. In a fair dice the long run expectation is that
‘2’ comes up 16.6…% of the time. Same here. So, your test is card 2 first versus any other card first with an expected value of .166…. Gene Maguin From: freek kuper [[hidden email]]
Hi First of all thanks for all the replies! I do have some experience with SPSS. But then I didn’t put the data in my self. I did a few simple tests, T-tests, regressions, put variables together etc., I think most of the basic stuff. I even had an exam about this. But now I have
to put the data in my self and find out what tests I should use. There is indeed a ranking of the cards plus the two follow up questions.
I’ll try to explain this sentence: “But I made a difference between search products (smartphone) and experience products (computer game). For both product categories I did exactly the same. Thus now I have two times an order of 6 ‘cards’ with two answers
each.” Every person first saw the six cards for the smartphone (thus the search product). I now have a ranking of the cards for the search product and the two follow up questions. Then I showed them the cards for the computer game (thus the experience
product). Now I have also a ranking of six cards with two follow up questions each. Thus now I have 30 responses for a search good and 30 responses (the same participants) for the experience product.
I hope this explains it. Now I want to see what the ranking is. In one of my hypotheses, based on scientific articles, I said that people will pick card 2 as the first choice. I want to check if this is significant (if that is even possible). But I might also be
able to this with just the descriptives. But then I also want to compare the rankings of the search goods with the rankings of the experience goods. To see if there is a difference. Another hypotheses of my is that people will choose for more positive reviews
for an experience good then for a search good. I hope I explained it a bit better now. Thanks again! Freek Verzonden vanuit
Mail voor Windows 10
I think I understand the basic design: Each person picks one of six cards and answers two questions about that card. Repeat that process for the five remaining cards. So your data is a ranking of the cards plus
the two followup questions about each card. However, I’m confused by this sentence, “But I made a difference between search products (smartphone) and experience products (computer game). For both product categories I did exactly
the same. Thus now I have two times an order of 6 ‘cards’ with two answers each.” Please explain this sentence in operational/behavioral terms. Your placement of this data value (variable 6) in the data layout you included
(Thank you very much for that) suggests that your people declared “Search” or “Experience” before you showed them the set of cards. So, the value of variable 6 applies to the set of responses that follow, i.e., variables 7-24. But, then you say this, “Thus
now I have two times an order of 6 ‘cards’ with two answers each.” Maybe there’s really a different story.
Confusion in à confusion out. Gene Maguin From: SPSSX(r) Discussion [[hidden email]]
On Behalf Of freek kuper Hi all! I’m stuck with my report. I collected some data, but now I don’t know what to do in SPSS. First some background about my report and data collection: I interviewed 30 people. They had to choose between 6 ‘cards’. Every card represent a product. This way I find out which product my participants would buy. If they choose one ‘card’, I asked two questions. One question about
credibility and one about where to buy this product. Both with a Likert scale of 1 to 5. I did this until there were no ‘cards’ left. Thus the results are something like this for example: Card 5, 3, 4; Card 1, 2, 2; Card 3, 4, 2; etc. till all six ‘cards’.
But I made a difference between search products (smartphone) and experience products (computer game). For both product categories I did exactly the same. Thus now I have two times an order of 6 ‘cards’ with two answers each.
I also asked their sex, age and three introduction questions with a Likert scale of 1 to 5 before I conducted the tests with the ‘cards’.
What I want to find out if there is a difference between the choices of cards that the participants made for search and experience products. Thus I want to compare them to each other. Do people make
other decisions for a search product than for an experience product. However I have no idea how to do it. My supervisor told me to use Anova in SPSS. I’ve checked some videos and manuals. But I don’t know how I can test, what I want to test. I put my data in SPSS in the following manner:
Variable 1 – Age Variable 2 – Sex Variable 3 – Introduction question 1 Variable 4 – Introduction question 2 Variable 5 – Introduction question 3 Variable 6 – Search product or Experience product Variable 7 – First choice (Thus did you choose card 1, 2, 3, 4, 5 or 6) Variable 8 – Credibility 1 Variable 9 – Where to buy (in a store or online) 1 Variable 10 – Second choice (Thus card 1 till 6 with exception of your first choice of course) Variable 11 – Credibility 2 Variable 12 – Where to buy 2 Variable 13 – Third choice (Thus card 1 till 6 with exception of the first two choices of course) Etc., till all six cards are gone. Thus what I want to find out is, if the order of cards people choose will be different for search and experience products, the credibility and where to buy this product for every card.
I wish you all a happy new year and I hope that one of you could help me out! Thanks in advance! Freek Verzonden vanuit
Mail voor Windows 10 ===================== To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text
except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD
===================== To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the command.
To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD
===================== To manage your subscription to SPSSX-L, send a message to
[hidden email]
(not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command
INFO REFCARD ===================== To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text
except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD ===================== To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD
===================== To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD
===================== To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD
|
| Free forum by Nabble | Edit this page |