Disclaimer: Novice user but great at following directions. Can use SPSS
successfully For my dissertation "The perceived resilience and stress of K-12 teachers with adverse childhood experiences", I surveyed teachers to collect yes/no to 15 questions on adverse childhood experiences, 10 likert questions on resilience, and 15 likert questions on stress. Also some basic demographics (age, years of teaching, urban or rural school setting). I used qualtrics to collect this information. I summed my 3 different sections (adverse, resilience, stress) in SPSS so now I have 3 variables with summed scores. I wanted to look at only the stress and resilience of teachers who reported 2 or more adverse experiences so I sorted it by <= 27 (yes is 1 point, so a perfect test was 30). Here are my questions: If I am wanting to establish correlations among these 3 variables, is summarizing them the best thing to do. If so, what statistical test should I perform? My hypothesis is: there is no statistically significant difference in resilience and stress of teachers with 2 or more adverse experiences. Also, when I want to throw in demographics for comparisons....how do I do that? -- Sent from: http://spssx-discussion.1045642.n5.nabble.com/ ===================== 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 |
Definitely, you want to discuss "the number of experiences", 1-15, which is
the number of YES responses - and not talk about a confusing score (going
the wrong direction) that runs from 15 to 30.
Similarly, though not as strongly, it is usually preferable to use the average, not the sum, for likert-type scoring, since the average lets you refer directly to the
verbal anchors. Name and score the Resilience and Stress variables in the direction
that makes sense for your discussion.
Without knowing the counts (sample, Experiences) and whether the means show much variance, it is impossible to frame the best presentation.
I would start with ... no, wait, I always start with doing a factor analysis on the items of any proposed scales, just to make sure that they all contribute in the same direction as I expected. (No "bad" items.) And I would do examine a scattergram of the two
averages to see the spread of scores and the full-sample correlation.
Then I would start with looking at the means of Resil and Stress for each number of
Experiences, that is, ANOVA by Experiences. If I wanted to form Groups out of the
counts of Experiences, and if Exper appears to be Poisson-distributed, I might use the boundaries based on the square-root of counts -- (0, 1, 2-4, 5-9, 10-15).
Finally - "My hypothesis is: there is no statistically significant difference in resilience and stress of teacherswith 2 or more adverse experiences."
Difference between whom? Difference in means according to the count of experiences?
"Is there a correlation between Exper and Stress or Resil, for those with 2+ experiences?"
(You say 2+, but your test with "27" implies 3+. Less confusion with actual counts, as I said.) - That would be a correlation on the selected sub-sample. (I'd probably use sqrt(Exper).)
Your other tests imply, probably, regression analyses.
-- Rich Ulrich From: SPSSX(r) Discussion <[hidden email]> on behalf of mandilogan <[hidden email]>
Sent: Thursday, December 7, 2017 6:17:33 PM To: [hidden email] Subject: How to approach questionnaire data Disclaimer: Novice user but great at following directions. Can use SPSS
successfully For my dissertation "The perceived resilience and stress of K-12 teachers with adverse childhood experiences", I surveyed teachers to collect yes/no to 15 questions on adverse childhood experiences, 10 likert questions on resilience, and 15 likert questions on stress. Also some basic demographics (age, years of teaching, urban or rural school setting). I used qualtrics to collect this information. I summed my 3 different sections (adverse, resilience, stress) in SPSS so now I have 3 variables with summed scores. I wanted to look at only the stress and resilience of teachers who reported 2 or more adverse experiences so I sorted it by <= 27 (yes is 1 point, so a perfect test was 30). Here are my questions: If I am wanting to establish correlations among these 3 variables, is summarizing them the best thing to do. If so, what statistical test should I perform? My hypothesis is: there is no statistically significant difference in resilience and stress of teachers with 2 or more adverse experiences. Also, when I want to throw in demographics for comparisons....how do I do that? -- Sent from: http://spssx-discussion.1045642.n5.nabble.com/ ===================== 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 |
I was trying to figure out how a 15 item list where each endorsement is 1 resulted in a 'perfect test' being 30.
Based on Rich's response it seems you've coded 1=Yes and 2=No. To do what Rich suggest, recode No=2 to No=0 so that your sum is the total number of experiences and the mean of each item (x100) is the % endorsing that item. Summing Likert scales comes to a completely different result if there are any missing values for an individual. Means are only appropriate if the items "hang together" well. It is highly unlikely that the stress items have high reliability (e.g., you wake up on the wrong side of the bed, your pet has been sick on the rug, the car won't start are all stressful but have none have anything to do with the others). And missing values will be an issue here too. Melissa ________________________________ From: SPSSX(r) Discussion <[hidden email]> on behalf of Rich Ulrich <[hidden email]> Sent: Thursday, December 7, 2017 7:24 PM To: [hidden email] Subject: Re: [SPSSX-L] How to approach questionnaire data Definitely, you want to discuss "the number of experiences", 1-15, which is the number of YES responses - and not talk about a confusing score (going the wrong direction) that runs from 15 to 30. Similarly, though not as strongly, it is usually preferable to use the average, not the sum, for likert-type scoring, since the average lets you refer directly to the verbal anchors. Name and score the Resilience and Stress variables in the direction that makes sense for your discussion. Without knowing the counts (sample, Experiences) and whether the means show much variance, it is impossible to frame the best presentation. I would start with ... no, wait, I always start with doing a factor analysis on the items of any proposed scales, just to make sure that they all contribute in the same direction as I expected. (No "bad" items.) And I would do examine a scattergram of the two averages to see the spread of scores and the full-sample correlation. Then I would start with looking at the means of Resil and Stress for each number of Experiences, that is, ANOVA by Experiences. If I wanted to form Groups out of the counts of Experiences, and if Exper appears to be Poisson-distributed, I might use the boundaries based on the square-root of counts -- (0, 1, 2-4, 5-9, 10-15). Finally - "My hypothesis is: there is no statistically significant difference in resilience and stress of teachers with 2 or more adverse experiences." Difference between whom? Difference in means according to the count of experiences? "Is there a correlation between Exper and Stress or Resil, for those with 2+ experiences?" (You say 2+, but your test with "27" implies 3+. Less confusion with actual counts, as I said.) - That would be a correlation on the selected sub-sample. (I'd probably use sqrt(Exper).) Your other tests imply, probably, regression analyses. -- Rich Ulrich ________________________________ From: SPSSX(r) Discussion <[hidden email]> on behalf of mandilogan <[hidden email]> Sent: Thursday, December 7, 2017 6:17:33 PM To: [hidden email] Subject: How to approach questionnaire data Disclaimer: Novice user but great at following directions. Can use SPSS successfully For my dissertation "The perceived resilience and stress of K-12 teachers with adverse childhood experiences", I surveyed teachers to collect yes/no to 15 questions on adverse childhood experiences, 10 likert questions on resilience, and 15 likert questions on stress. Also some basic demographics (age, years of teaching, urban or rural school setting). I used qualtrics to collect this information. I summed my 3 different sections (adverse, resilience, stress) in SPSS so now I have 3 variables with summed scores. I wanted to look at only the stress and resilience of teachers who reported 2 or more adverse experiences so I sorted it by <= 27 (yes is 1 point, so a perfect test was 30). Here are my questions: If I am wanting to establish correlations among these 3 variables, is summarizing them the best thing to do. If so, what statistical test should I perform? My hypothesis is: there is no statistically significant difference in resilience and stress of teachers with 2 or more adverse experiences. Also, when I want to throw in demographics for comparisons....how do I do that? -- Sent from: http://spssx-discussion.1045642.n5.nabble.com/ ===================== 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]<mailto:[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 ________________________________ This correspondence contains proprietary information some or all of which may be legally privileged; it is for the intended recipient only. If you are not the intended recipient you must not use, disclose, distribute, copy, print, or rely on this correspondence and completely dispose of the correspondence immediately. Please notify the sender if you have received this email in error. NOTE: Messages to or from the State of Connecticut domain may be subject to the Freedom of Information statutes and regulations. ===================== 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 |
As a beginner, you may need assistance at a more basic level. Melissa is right to advise recoding your data Without seeing your data (and assuming variable names) I would have started with something like: [All syntax unverified] COUNT adverse = ad1 to ad15 (1) /resil =res1 to res10 (1) /stress = str1 to stre15 (1). FORMAT adverse to stress (f2.0). FREQ adverse to stress /his nor. However COUNT yields a score even if items are missing, so more accurate scores would be obtained by: RECODE ad1 to ad15 resil1 to resil10 str1 to str15 (2 = 0) . *Subtract no of items to yield ratio scale with true 0. COMPUTE adverse = sum.15 (ad1 to ad15) – 15. COMPUTE resil = sum.10 (resil1 to resil10) -10. COMPUTE stress = sum.15 (stress1 to stress15) – 15. FORMATS adverse to stress (f2.0). FREQ adverse to stress /his nor. RECODE adverse (2 thru 15 =2) into adversegrp. FORMATS adversegrp (f2.0). VAR LAB adversegrp 'No of adverse experiences'. VAL LAB adversegrp 0 'None' 1 'One' 2 'Two or more'. FREQ adversegrp. MEANS resil stress by adversegrp. How many cases do you have? A useful trick with binary data is to use MULT RESP. MULT RESP groups resilgrp (res1 to res10 (1)) stressgrp (str1 to str15 (1)) /freq resilgrp stressgrp /tab resilgrp by resilgrp /tab stressgrp by stressgrp. There are (very) basic tutorials on simple attitude scoring on http://surveyresearch.weebly.com/352-teenage-attitudes-tutorials.html Hope this helps. John F Hall [Retired academic survey researcher] IBM-SPSS Academic Author 9900074 Email: [hidden email] Website: http://surveyresearch.weebly.com/ SPSS course: http://surveyresearch.weebly.com/1-survey-analysis-workshop-spss.html Research: http://surveyresearch.weebly.com/3-subjective-social-indicators-quality-of-life.html -----Original Message----- I was trying to figure out how a 15 item list where each endorsement is 1 resulted in a 'perfect test' being 30. Based on Rich's response it seems you've coded 1=Yes and 2=No. To do what Rich suggest, recode No=2 to No=0 so that your sum is the total number of experiences and the mean of each item (x100) is the % endorsing that item. Summing Likert scales comes to a completely different result if there are any missing values for an individual. Means are only appropriate if the items "hang together" well. It is highly unlikely that the stress items have high reliability (e.g., you wake up on the wrong side of the bed, your pet has been sick on the rug, the car won't start are all stressful but have none have anything to do with the others). And missing values will be an issue here too. Melissa ________________________________ From: SPSSX(r) Discussion <[hidden email]> on behalf of Rich Ulrich <[hidden email]> Sent: Thursday, December 7, 2017 7:24 PM To: [hidden email] Subject: Re: [SPSSX-L] How to approach questionnaire data Definitely, you want to discuss "the number of experiences", 1-15, which is the number of YES responses - and not talk about a confusing score (going the wrong direction) that runs from 15 to 30. Similarly, though not as strongly, it is usually preferable to use the average, not the sum, for likert-type scoring, since the average lets you refer directly to the verbal anchors. Name and score the Resilience and Stress variables in the direction that makes sense for your discussion. Without knowing the counts (sample, Experiences) and whether the means show much variance, it is impossible to frame the best presentation. I would start with ... no, wait, I always start with doing a factor analysis on the items of any proposed scales, just to make sure that they all contribute in the same direction as I expected. (No "bad" items.) And I would do examine a scattergram of the two averages to see the spread of scores and the full-sample correlation. Then I would start with looking at the means of Resil and Stress for each number of Experiences, that is, ANOVA by Experiences. If I wanted to form Groups out of the counts of Experiences, and if Exper appears to be Poisson-distributed, I might use the boundaries based on the square-root of counts -- (0, 1, 2-4, 5-9, 10-15). Finally - "My hypothesis is: there is no statistically significant difference in resilience and stress of teachers with 2 or more adverse experiences." Difference between whom? Difference in means according to the count of experiences? "Is there a correlation between Exper and Stress or Resil, for those with 2+ experiences?" (You say 2+, but your test with "27" implies 3+. Less confusion with actual counts, as I said.) - That would be a correlation on the selected sub-sample. (I'd probably use sqrt(Exper).) Your other tests imply, probably, regression analyses. -- Rich Ulrich ________________________________ From: SPSSX(r) Discussion <[hidden email]> on behalf of mandilogan <[hidden email]> Sent: Thursday, December 7, 2017 6:17:33 PM To: [hidden email] Subject: How to approach questionnaire data Disclaimer: Novice user but great at following directions. Can use SPSS successfully For my dissertation "The perceived resilience and stress of K-12 teachers with adverse childhood experiences", I surveyed teachers to collect yes/no to 15 questions on adverse childhood experiences, 10 likert questions on resilience, and 15 likert questions on stress. Also some basic demographics (age, years of teaching, urban or rural school setting). I used qualtrics to collect this information. I summed my 3 different sections (adverse, resilience, stress) in SPSS so now I have 3 variables with summed scores. I wanted to look at only the stress and resilience of teachers who reported 2 or more adverse experiences so I sorted it by <= 27 (yes is 1 point, so a perfect test was 30). Here are my questions: If I am wanting to establish correlations among these 3 variables, is summarizing them the best thing to do. If so, what statistical test should I perform? My hypothesis is: there is no statistically significant difference in resilience and stress of teachers with 2 or more adverse experiences. Also, when I want to throw in demographics for comparisons....how do I do that? -- Sent from: http://spssx-discussion.1045642.n5.nabble.com/ ===================== 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 ________________________________ This correspondence contains proprietary information some or all of which may be legally privileged; it is for the intended recipient only. If you are not the intended recipient you must not use, disclose, distribute, copy, print, or rely on this correspondence and completely dispose of the correspondence immediately. Please notify the sender if you have received this email in error. NOTE: Messages to or from the State of Connecticut domain may be subject to the Freedom of Information statutes and regulations. ===================== 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 |