How to code hazards across house rooms
Locked
 
|
Dear List: I have just received an SPSS data file from a colleague that contains responses from a City’s elderly residents who reported on safety hazards within their homes/apartments. Each room could have anywhere from 10 to 14 hazards that the residents had to check off. We are interested in getting a total hazards score for the five rooms (across the kitchen, living room, bedroom, dining room, and bathroom). However, I have noticed that not all of the residents had five rooms. Some lived in efficiency apartments and only had two rooms. If I totaled up the number of hazards across all rooms those with fewer rooms would automatically have lower hazards. If I obtained the mean across all of the rooms than residents could have the same mean number of hazards but it would be based upon a different number of rooms. I am trying to figure out how to factor in the number of rooms. Some way of weighting the scores but I am a bit dumbfounded on this. I realize that this is not an SPSS question but would appreciate suggestions (if you would like to chime in) as I begin to figure out how to analyze the data. Thanks in advance, martin Martin F. Sherman, Ph.D. Professor of Psychology Director of Masters Education in Psychology: Thesis Track Loyola University Maryland Department of Psychology 222 B Beatty Hall 4501 North Charles Street Baltimore, MD 21210 410-617-2417 |
Re: How to code hazards across house rooms
|
Martin What does the data file actually look like? How many rows of data are there per residence? Is each room type in a separate data column or on a separate data row? Are the hazards coded on the same row? Does each hazard have a separate column or is there just a list of codes regardless of column? You mention up to five room types: do any dwellings have more than one of each? Sounds to me as if you need look at MULT RESPONSE to analyse the lists of hazards overall and for each room type. You can also use COUNT to create several indices by counting the number of different hazards in each room type, or the same hazard across all room types. If you send me, in complete confidence and off-list, a copy of the original questionnaire/schedule and (an extract of) the Data Editor, I may be able to give more detailed help. John F Hall (Mr) [retired academic survey researcher] Email: [hidden email] Website: www.surveyresearch.weebly.com From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Martin Sherman Dear List: I have just received an SPSS data file from a colleague that contains responses from a City’s elderly residents who reported on safety hazards within their homes/apartments. Each room could have anywhere from 10 to 14 hazards that the residents had to check off. We are interested in getting a total hazards score for the five rooms (across the kitchen, living room, bedroom, dining room, and bathroom). However, I have noticed that not all of the residents had five rooms. Some lived in efficiency apartments and only had two rooms. If I totaled up the number of hazards across all rooms those with fewer rooms would automatically have lower hazards. If I obtained the mean across all of the rooms than residents could have the same mean number of hazards but it would be based upon a different number of rooms. I am trying to figure out how to factor in the number of rooms. Some way of weighting the scores but I am a bit dumbfounded on this. I realize that this is not an SPSS question but would appreciate suggestions (if you would like to chime in) as I begin to figure out how to analyze the data. Thanks in advance, martin Martin F. Sherman, Ph.D. Professor of Psychology Director of Masters Education in Psychology: Thesis Track Loyola University Maryland Department of Psychology 222 B Beatty Hall 4501 North Charles Street Baltimore, MD 21210 410-617-2417 |
Re: How to code hazards across house rooms
|
|
In reply to this post by msherman
Think about what someone will be writing up. What is the narrative?
(You don't tell us that, but you might assume that it is safer to look at a wider rather than a narrower range of options, even if there is a firmly stated, and limited, purpose.) It seems to me that collapsing everything you know about hazards into one number is premature. Just for starters, you *will* want to know the total number of hazards, which requires the the total number of all hazards in each household -- which might be more than 14. Hazards-per-room also seems useful to report. The "number of different hazards" might go along with a mention of non-prevalent hazards, and gives an additional perspective on risk. That is the one that might be important for assessing the success of any educational or other preventative measures. -- Rich Ulrich Date: Thu, 7 Feb 2013 19:21:43 -0500 From: [hidden email] Subject: How to code hazards across house rooms To: [hidden email] Dear List: I have just received an SPSS data file from a colleague that contains responses from a City’s elderly residents who reported on safety hazards within their homes/apartments. Each room could have anywhere from 10 to 14 hazards that the residents had to check off. We are interested in getting a total hazards score for the five rooms (across the kitchen, living room, bedroom, dining room, and bathroom). However, I have noticed that not all of the residents had five rooms. Some lived in efficiency apartments and only had two rooms. If I totaled up the number of hazards across all rooms those with fewer rooms would automatically have lower hazards. If I obtained the mean across all of the rooms than residents could have the same mean number of hazards but it would be based upon a different number of rooms. I am trying to figure out how to factor in the number of rooms. Some way of weighting the scores but I am a bit dumbfounded on this. I realize that this is not an SPSS question but would appreciate suggestions (if you would like to chime in) as I begin to figure out how to analyze the data. Thanks in advance, martin
... |
Re: How to code hazards across house rooms
|
|
Someone else has already suggested Hazards per Room.
If such data is provided, Hazards per Square Foot of space might help you get around the number of rooms issue. ... Mark Miller
On Thu, Feb 7, 2013 at 11:18 PM, Rich Ulrich <[hidden email]> wrote:
|
Re: How to code hazards across house rooms
|
|
In reply to this post by msherman
Would it make sense to look at percentage vs. number of hazards? So, for each resident, the outcome would be the percentage of possible hazards they experienced, which might circumvent the number of rooms issue. Todd Zoblotsky From: SPSSX(r) Discussion [mailto:[hidden email]]
On Behalf Of Martin Sherman Dear List: I have just received an SPSS data file from a colleague that contains responses from a City’s elderly residents who reported on safety hazards within their homes/apartments. Each room could have anywhere from 10 to 14 hazards
that the residents had to check off. We are interested in getting a total hazards score for the five rooms (across the kitchen, living room, bedroom, dining room, and bathroom). However, I have noticed that not all of the residents had five rooms. Some lived
in efficiency apartments and only had two rooms. If I totaled up the number of hazards across all rooms those with fewer rooms would automatically have lower hazards. If I obtained the mean across all of the rooms than residents could have the same mean number
of hazards but it would be based upon a different number of rooms. I am trying to figure out how to factor in the number of rooms. Some way of weighting the scores but I am a bit dumbfounded on this. I realize that this is not an SPSS question but would appreciate
suggestions (if you would like to chime in) as I begin to figure out how to analyze the data. Thanks in advance, martin Martin F. Sherman, Ph.D. Professor of Psychology Director of Masters Education in Psychology: Thesis Track Loyola University Maryland Department of Psychology 222 B Beatty Hall 4501 North Charles Street Baltimore, MD 21210 410-617-2417 |
Re: How to code hazards across house rooms
|
|
In reply to this post by msherman
Martin, I addition to what others have said, I think that you might want to consider multiple ways of working with the data. Certainly compute the mean, hazards per room. I don’t know that weighting is the correct concept here because weighting seems to be part of the design of a complex survey design. I think it might be useful to investigate total hazards as a function of number of rooms. I wonder if the relationship is linear. I could imagine that it’s not because number of rooms probably has some correlation magnitude with wealth/income, which might well be related to hazard opportunity. Depending on how you asked the questions, a hazard is an unremediated hazard opportunity. Gene Maguin From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Martin Sherman Dear List: I have just received an SPSS data file from a colleague that contains responses from a City’s elderly residents who reported on safety hazards within their homes/apartments. Each room could have anywhere from 10 to 14 hazards that the residents had to check off. We are interested in getting a total hazards score for the five rooms (across the kitchen, living room, bedroom, dining room, and bathroom). However, I have noticed that not all of the residents had five rooms. Some lived in efficiency apartments and only had two rooms. If I totaled up the number of hazards across all rooms those with fewer rooms would automatically have lower hazards. If I obtained the mean across all of the rooms than residents could have the same mean number of hazards but it would be based upon a different number of rooms. I am trying to figure out how to factor in the number of rooms. Some way of weighting the scores but I am a bit dumbfounded on this. I realize that this is not an SPSS question but would appreciate suggestions (if you would like to chime in) as I begin to figure out how to analyze the data. Thanks in advance, martin Martin F. Sherman, Ph.D. Professor of Psychology Director of Masters Education in Psychology: Thesis Track Loyola University Maryland Department of Psychology 222 B Beatty Hall 4501 North Charles Street Baltimore, MD 21210 410-617-2417 |
|
|
In reply to this post by John F Hall
John: The data file consists of one row for each resident and each resident has all of the variables. When a resident is missing a room a sysmis was used. Using count or multiple response will give me the number of hazards per room (and total) but again my concern is about the residents that do not have all five rooms (and have sysmis for those missing hazards per room). martin From: John F Hall [mailto:[hidden email]] Martin What does the data file actually look like? How many rows of data are there per residence? Is each room type in a separate data column or on a separate data row? Are the hazards coded on the same row? Does each hazard have a separate column or is there just a list of codes regardless of column? You mention up to five room types: do any dwellings have more than one of each? Sounds to me as if you need look at MULT RESPONSE to analyse the lists of hazards overall and for each room type. You can also use COUNT to create several indices by counting the number of different hazards in each room type, or the same hazard across all room types. If you send me, in complete confidence and off-list, a copy of the original questionnaire/schedule and (an extract of) the Data Editor, I may be able to give more detailed help. John F Hall (Mr) [retired academic survey researcher] Email: [hidden email] Website: www.surveyresearch.weebly.com From: SPSSX(r) Discussion [[hidden email]] On Behalf Of Martin Sherman Dear List: I have just received an SPSS data file from a colleague that contains responses from a City’s elderly residents who reported on safety hazards within their homes/apartments. Each room could have anywhere from 10 to 14 hazards that the residents had to check off. We are interested in getting a total hazards score for the five rooms (across the kitchen, living room, bedroom, dining room, and bathroom). However, I have noticed that not all of the residents had five rooms. Some lived in efficiency apartments and only had two rooms. If I totaled up the number of hazards across all rooms those with fewer rooms would automatically have lower hazards. If I obtained the mean across all of the rooms than residents could have the same mean number of hazards but it would be based upon a different number of rooms. I am trying to figure out how to factor in the number of rooms. Some way of weighting the scores but I am a bit dumbfounded on this. I realize that this is not an SPSS question but would appreciate suggestions (if you would like to chime in) as I begin to figure out how to analyze the data. Thanks in advance, martin Martin F. Sherman, Ph.D. Professor of Psychology Director of Masters Education in Psychology: Thesis Track Loyola University Maryland Department of Psychology 222 B Beatty Hall 4501 North Charles Street Baltimore, MD 21210 410-617-2417 |
Re: How to code hazards across house rooms
|
|
I like the percentage idea. Â What I keep thinking about is that if a person doesn't have a dining room, they do the things they would do in the dining room in the kitchen. Â If they don't have a bedroom, they do the things they would do in the bedroom in the living room. Â They still spend the same number of hours in their apartment. Â To some extent it makes conceptual sense to weight the rooms by the number of functions they perform. Â But I don't know that you could justify that rigorously.
On Fri, Feb 8, 2013 at 8:49 AM, Martin Sherman <[hidden email]> wrote:
This e-mail transmission and any attachments accompanying it may contain confidential and/or proprietary information and is intended only for the person or entity to whom it was originally addressed. If you are not the intended recipient, you are hereby notified that any disclosure, copying, distribution or other use of this information is strictly prohibited. Any unauthorized interception of this transmission is illegal. If you have received this transmission in error, please notify the sender by reply e-mail, and then destroy all copies of this transmission
|
|
|
Martha: Thanks for the response. I did not think about the amount of time spent in the residence would basically be equated regardless of the number of rooms. Getting the average number of hazards across the rooms (equated all for five rooms) does that but I still worry about the sheer number. That is, it might be that only certain types of hazards are more likely in certain rooms than in others. Getting the average prorated across five rooms misses out on that. martin From: Martha Hewett [mailto:[hidden email]] I like the percentage idea. What I keep thinking about is that if a person doesn't have a dining room, they do the things they would do in the dining room in the kitchen. If they don't have a bedroom, they do the things they would do in the bedroom in the living room. They still spend the same number of hours in their apartment. To some extent it makes conceptual sense to weight the rooms by the number of functions they perform. But I don't know that you could justify that rigorously.
On Fri, Feb 8, 2013 at 8:49 AM, Martin Sherman <[hidden email]> wrote: John: The data file consists of one row for each resident and each resident has all of the variables. When a resident is missing a room a sysmis was used. Using count or multiple response will give me the number of hazards per room (and total) but again my concern is about the residents that do not have all five rooms (and have sysmis for those missing hazards per room). martin From: John F Hall [mailto:[hidden email]] Martin What does the data file actually look like? How many rows of data are there per residence? Is each room type in a separate data column or on a separate data row? Are the hazards coded on the same row? Does each hazard have a separate column or is there just a list of codes regardless of column? You mention up to five room types: do any dwellings have more than one of each? Sounds to me as if you need look at MULT RESPONSE to analyse the lists of hazards overall and for each room type. You can also use COUNT to create several indices by counting the number of different hazards in each room type, or the same hazard across all room types. If you send me, in complete confidence and off-list, a copy of the original questionnaire/schedule and (an extract of) the Data Editor, I may be able to give more detailed help. John F Hall (Mr) [retired academic survey researcher] Email: [hidden email] Website: www.surveyresearch.weebly.com From: SPSSX(r) Discussion [[hidden email]] On Behalf Of Martin Sherman
Dear List: I have just received an SPSS data file from a colleague that contains responses from a City’s elderly residents who reported on safety hazards within their homes/apartments. Each room could have anywhere from 10 to 14 hazards that the residents had to check off. We are interested in getting a total hazards score for the five rooms (across the kitchen, living room, bedroom, dining room, and bathroom). However, I have noticed that not all of the residents had five rooms. Some lived in efficiency apartments and only had two rooms. If I totaled up the number of hazards across all rooms those with fewer rooms would automatically have lower hazards. If I obtained the mean across all of the rooms than residents could have the same mean number of hazards but it would be based upon a different number of rooms. I am trying to figure out how to factor in the number of rooms. Some way of weighting the scores but I am a bit dumbfounded on this. I realize that this is not an SPSS question but would appreciate suggestions (if you would like to chime in) as I begin to figure out how to analyze the data. Thanks in advance, martin Martin F. Sherman, Ph.D. Professor of Psychology Director of Masters Education in Psychology: Thesis Track Loyola University Maryland Department of Psychology 222 B Beatty Hall 4501 North Charles Street Baltimore, MD 21210 410-617-2417 This e-mail transmission and any attachments accompanying it may contain confidential and/or proprietary information and is intended only for the person or entity to whom it was originally addressed. If you are not the intended recipient, you are hereby notified that any disclosure, copying, distribution or other use of this information is strictly prohibited. Any unauthorized interception of this transmission is illegal. If you have received this transmission in error, please notify the sender by reply e-mail, and then destroy all copies of this transmission |
Re: How to code hazards across house rooms
|
|
Further to what Eugene Maguin said, I think that you should report results both as means and totals per room and per house/flat as appropriate. Arguably, a combined kitchen/living room may be less hazardous in itself than the separate rooms, because one is more likely to see the hazard activate, i.e. the fewer the rooms the more likely you are to see and take action sooner when necessary. I suggest you would then discuss the different viewpoints in your analysis rather than just trying to shoehorn everything into one classification.
|
Re: How to code hazards across house rooms
|
|
In reply to this post by Maguin, Eugene
Yes, I am. Include rooms. It depends on how you are going to analyze the data. Re: hazard opportunity. I don’t know anything about this topic beyond just lived experiences. I don’t know what’s on your hazards list but let’s say extension cords. So, plausibly, an extension cord laying on the floor where people walk is a hazard. So, suppose I tape it down. Still a hazard? Maybe. Harder to trip on but still possible. But it’s also been at least partly remediated. Gene Maguin From: Martin Sherman [mailto:[hidden email]] Gene: thanks. Are you suggesting beside a total hazards variable that I include a number of rooms variable. It is possible then that the total number of hazards might be highly correlated with the number of rooms. And are you saying that a hazard is a hazard (unremediated hazard opportunity). martin From: SPSSX(r) Discussion [[hidden email]] On Behalf Of Maguin, Eugene Martin, I addition to what others have said, I think that you might want to consider multiple ways of working with the data. Certainly compute the mean, hazards per room. I don’t know that weighting is the correct concept here because weighting seems to be part of the design of a complex survey design. I think it might be useful to investigate total hazards as a function of number of rooms. I wonder if the relationship is linear. I could imagine that it’s not because number of rooms probably has some correlation magnitude with wealth/income, which might well be related to hazard opportunity. Depending on how you asked the questions, a hazard is an unremediated hazard opportunity. Gene Maguin From: SPSSX(r) Discussion [[hidden email]] On Behalf Of Martin Sherman Dear List: I have just received an SPSS data file from a colleague that contains responses from a City’s elderly residents who reported on safety hazards within their homes/apartments. Each room could have anywhere from 10 to 14 hazards that the residents had to check off. We are interested in getting a total hazards score for the five rooms (across the kitchen, living room, bedroom, dining room, and bathroom). However, I have noticed that not all of the residents had five rooms. Some lived in efficiency apartments and only had two rooms. If I totaled up the number of hazards across all rooms those with fewer rooms would automatically have lower hazards. If I obtained the mean across all of the rooms than residents could have the same mean number of hazards but it would be based upon a different number of rooms. I am trying to figure out how to factor in the number of rooms. Some way of weighting the scores but I am a bit dumbfounded on this. I realize that this is not an SPSS question but would appreciate suggestions (if you would like to chime in) as I begin to figure out how to analyze the data. Thanks in advance, martin Martin F. Sherman, Ph.D. Professor of Psychology Director of Masters Education in Psychology: Thesis Track Loyola University Maryland Department of Psychology 222 B Beatty Hall 4501 North Charles Street Baltimore, MD 21210 410-617-2417 |
Re: How to code hazards across house rooms
|
|
In reply to this post by msherman
Martin Do you have a variable; [Number of rooms]? Do you have a Yes/No code for each room type? If room type is coded 1/0 you could do something like: COUNT rooms = living, kitchen, dining, bed, bath (1). If it’s coded by type (eg 1 – 5) then: Count rooms = living, kitchen, dining, bed, bath (1 thru 5). . . . or whatever your room variables are called. Assuming you have 1 column for each room type and 5 x 14 columns for the hazard types, and assuming the same coding list applies to each room type, you should have variables: [my names] Livhaz1-14, kitchhaz 1-14, dinhaz1-14, bedhaz 1-14, bathhaz1-14 These may be coded 1,0 + sysmis or 1-14 + sysmis. If the latter then something like: Mult resp groups = roomtype (living, kitchen, dining, bed, bath (1,5)) /hazardtype = (Livhaz1 to livhaz14, kitchhaz1 to kitchhaz14, dinhaz1 to dinhaz14, Bedhaz1 to bedhaz14, bathhaz1 to bathhaz14 (1,14)) /freq roomtype hazardtype /tables = hazardtype by roomtype /cel cou col. You can also get hazard type for each room by something like: Mult resp groups Khaz, (kitchhaz1 to kitchhaz14 (1,14)) /dhaz (dinhaz1 to dhaz14 (1,14)) /lhaz (livhaz1 to livhaz14 (1,14)) /behaz(bedhaz1 to bedhaz14 (1,14)) /bahaz (bathhaz1 to bathhaz14 (1,14)) /freq khaz to bahaz. If your room types and hazard types are coded 0/1 you can run mult resp in dichotomous mode: just replace (1,14) with (1) in the above (all untested). John F Hall (Mr) [retired academic survey researcher] Email: [hidden email] Website: www.surveyresearch.weebly.com From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Martin Sherman Martha: Thanks for the response. I did not think about the amount of time spent in the residence would basically be equated regardless of the number of rooms. Getting the average number of hazards across the rooms (equated all for five rooms) does that but I still worry about the sheer number. That is, it might be that only certain types of hazards are more likely in certain rooms than in others. Getting the average prorated across five rooms misses out on that. martin From: Martha Hewett [[hidden email]] I like the percentage idea. What I keep thinking about is that if a person doesn't have a dining room, they do the things they would do in the dining room in the kitchen. If they don't have a bedroom, they do the things they would do in the bedroom in the living room. They still spend the same number of hours in their apartment. To some extent it makes conceptual sense to weight the rooms by the number of functions they perform. But I don't know that you could justify that rigorously.
On Fri, Feb 8, 2013 at 8:49 AM, Martin Sherman <[hidden email]> wrote: John: The data file consists of one row for each resident and each resident has all of the variables. When a resident is missing a room a sysmis was used. Using count or multiple response will give me the number of hazards per room (and total) but again my concern is about the residents that do not have all five rooms (and have sysmis for those missing hazards per room). martin From: John F Hall [mailto:[hidden email]] Martin What does the data file actually look like? How many rows of data are there per residence? Is each room type in a separate data column or on a separate data row? Are the hazards coded on the same row? Does each hazard have a separate column or is there just a list of codes regardless of column? You mention up to five room types: do any dwellings have more than one of each? Sounds to me as if you need look at MULT RESPONSE to analyse the lists of hazards overall and for each room type. You can also use COUNT to create several indices by counting the number of different hazards in each room type, or the same hazard across all room types. If you send me, in complete confidence and off-list, a copy of the original questionnaire/schedule and (an extract of) the Data Editor, I may be able to give more detailed help. John F Hall (Mr) [retired academic survey researcher] Email: [hidden email] Website: www.surveyresearch.weebly.com From: SPSSX(r) Discussion [[hidden email]] On Behalf Of Martin Sherman
Dear List: I have just received an SPSS data file from a colleague that contains responses from a City’s elderly residents who reported on safety hazards within their homes/apartments. Each room could have anywhere from 10 to 14 hazards that the residents had to check off. We are interested in getting a total hazards score for the five rooms (across the kitchen, living room, bedroom, dining room, and bathroom). However, I have noticed that not all of the residents had five rooms. Some lived in efficiency apartments and only had two rooms. If I totaled up the number of hazards across all rooms those with fewer rooms would automatically have lower hazards. If I obtained the mean across all of the rooms than residents could have the same mean number of hazards but it would be based upon a different number of rooms. I am trying to figure out how to factor in the number of rooms. Some way of weighting the scores but I am a bit dumbfounded on this. I realize that this is not an SPSS question but would appreciate suggestions (if you would like to chime in) as I begin to figure out how to analyze the data. Thanks in advance, martin Martin F. Sherman, Ph.D. Professor of Psychology Director of Masters Education in Psychology: Thesis Track Loyola University Maryland Department of Psychology 222 B Beatty Hall 4501 North Charles Street Baltimore, MD 21210 410-617-2417 This e-mail transmission and any attachments accompanying it may contain confidential and/or proprietary information and is intended only for the person or entity to whom it was originally addressed. If you are not the intended recipient, you are hereby notified that any disclosure, copying, distribution or other use of this information is strictly prohibited. Any unauthorized interception of this transmission is illegal. If you have received this transmission in error, please notify the sender by reply e-mail, and then destroy all copies of this transmission |
|
|
John: what we have is the following. LivingRoomHaz1 (Poor lighting) 1=Yes, 2=no LivingRoomHaz2 (Clutter) 1=Yes, 2= no LivingRoomHaz3 (Cockroaches) 1 =Yes, 2 = no Etc LivingRoomHaz13 (smoking materials) 1=Yes, 2=No total of 13 hazards for Living Room BedRoomHaz1 (Poor lighting) 1 = Yes, 2 = No Etc BedRoomHaz14 Total of 14 hazards for Bedroom KitchenHaz1 (poor lighting) Etc KitchenHaz14 Total of 14 hazards for Kitchen BathroomHaz1 Etc BathroomHaz14 Total of 14 hazards for Bathroom HallwayHaz1 Etc HallwayHaz15 Total of 15 hazards for Hallway GeneralHaz1 Etc GeneralHaz8 Total of 8 hazards for General. Total Number of Hazards = 78. From: John F Hall [mailto:[hidden email]] Martin Do you have a variable; [Number of rooms]? Do you have a Yes/No code for each room type? If room type is coded 1/0 you could do something like: COUNT rooms = living, kitchen, dining, bed, bath (1). If it’s coded by type (eg 1 – 5) then: Count rooms = living, kitchen, dining, bed, bath (1 thru 5). . . . or whatever your room variables are called. Assuming you have 1 column for each room type and 5 x 14 columns for the hazard types, and assuming the same coding list applies to each room type, you should have variables: [my names] Livhaz1-14, kitchhaz 1-14, dinhaz1-14, bedhaz 1-14, bathhaz1-14 These may be coded 1,0 + sysmis or 1-14 + sysmis. If the latter then something like: Mult resp groups = roomtype (living, kitchen, dining, bed, bath (1,5)) /hazardtype = (Livhaz1 to livhaz14, kitchhaz1 to kitchhaz14, dinhaz1 to dinhaz14, Bedhaz1 to bedhaz14, bathhaz1 to bathhaz14 (1,14)) /freq roomtype hazardtype /tables = hazardtype by roomtype /cel cou col. You can also get hazard type for each room by something like: Mult resp groups Khaz, (kitchhaz1 to kitchhaz14 (1,14)) /dhaz (dinhaz1 to dhaz14 (1,14)) /lhaz (livhaz1 to livhaz14 (1,14)) /behaz(bedhaz1 to bedhaz14 (1,14)) /bahaz (bathhaz1 to bathhaz14 (1,14)) /freq khaz to bahaz. If your room types and hazard types are coded 0/1 you can run mult resp in dichotomous mode: just replace (1,14) with (1) in the above (all untested). John F Hall (Mr) [retired academic survey researcher] Email: [hidden email] Website: www.surveyresearch.weebly.com From: SPSSX(r) Discussion [[hidden email]] On Behalf Of Martin Sherman Martha: Thanks for the response. I did not think about the amount of time spent in the residence would basically be equated regardless of the number of rooms. Getting the average number of hazards across the rooms (equated all for five rooms) does that but I still worry about the sheer number. That is, it might be that only certain types of hazards are more likely in certain rooms than in others. Getting the average prorated across five rooms misses out on that. martin From: Martha Hewett [[hidden email]] I like the percentage idea. What I keep thinking about is that if a person doesn't have a dining room, they do the things they would do in the dining room in the kitchen. If they don't have a bedroom, they do the things they would do in the bedroom in the living room. They still spend the same number of hours in their apartment. To some extent it makes conceptual sense to weight the rooms by the number of functions they perform. But I don't know that you could justify that rigorously.
On Fri, Feb 8, 2013 at 8:49 AM, Martin Sherman <[hidden email]> wrote: John: The data file consists of one row for each resident and each resident has all of the variables. When a resident is missing a room a sysmis was used. Using count or multiple response will give me the number of hazards per room (and total) but again my concern is about the residents that do not have all five rooms (and have sysmis for those missing hazards per room). martin From: John F Hall [mailto:[hidden email]] Martin What does the data file actually look like? How many rows of data are there per residence? Is each room type in a separate data column or on a separate data row? Are the hazards coded on the same row? Does each hazard have a separate column or is there just a list of codes regardless of column? You mention up to five room types: do any dwellings have more than one of each? Sounds to me as if you need look at MULT RESPONSE to analyse the lists of hazards overall and for each room type. You can also use COUNT to create several indices by counting the number of different hazards in each room type, or the same hazard across all room types. If you send me, in complete confidence and off-list, a copy of the original questionnaire/schedule and (an extract of) the Data Editor, I may be able to give more detailed help. John F Hall (Mr) [retired academic survey researcher] Email: [hidden email] Website: www.surveyresearch.weebly.com From: SPSSX(r) Discussion [[hidden email]] On Behalf Of Martin Sherman
Dear List: I have just received an SPSS data file from a colleague that contains responses from a City’s elderly residents who reported on safety hazards within their homes/apartments. Each room could have anywhere from 10 to 14 hazards that the residents had to check off. We are interested in getting a total hazards score for the five rooms (across the kitchen, living room, bedroom, dining room, and bathroom). However, I have noticed that not all of the residents had five rooms. Some lived in efficiency apartments and only had two rooms. If I totaled up the number of hazards across all rooms those with fewer rooms would automatically have lower hazards. If I obtained the mean across all of the rooms than residents could have the same mean number of hazards but it would be based upon a different number of rooms. I am trying to figure out how to factor in the number of rooms. Some way of weighting the scores but I am a bit dumbfounded on this. I realize that this is not an SPSS question but would appreciate suggestions (if you would like to chime in) as I begin to figure out how to analyze the data. Thanks in advance, martin Martin F. Sherman, Ph.D. Professor of Psychology Director of Masters Education in Psychology: Thesis Track Loyola University Maryland Department of Psychology 222 B Beatty Hall 4501 North Charles Street Baltimore, MD 21210 410-617-2417 This e-mail transmission and any attachments accompanying it may contain confidential and/or proprietary information and is intended only for the person or entity to whom it was originally addressed. If you are not the intended recipient, you are hereby notified that any disclosure, copying, distribution or other use of this information is strictly prohibited. Any unauthorized interception of this transmission is illegal. If you have received this transmission in error, please notify the sender by reply e-mail, and then destroy all copies of this transmission |
| Free forum by Nabble | Edit this page |