My research aims to assess whether parents who have been involved in a relationship with a social worker are at higher risk of child abuse/neglect.

I am trying to establish whether there exists a relationship between social workers' attitude toward the parents that reflect an overall outcome of the case (positive or negative). I want to assess whether social worker attitudes have an effect on parental responses.

  • The questionnaire was only given to parents who have a relationship with a social worker.
  • There were 20 participants (16 females and 4 males).
  • I aggregated my 13 Likert items/questions into a single attitude (word) per each.
  • I also reversed the order of the Likert scale when I entered the data, so that a score of 5 stands for "strongly disagree", whereas a score of 1 reads "strongly agree". This was done because of the way my Likert items were asked.


  • Can I run a test between genders given that there are an uneven number of males and females?
  • What analyses can I run to test my research question?
  • 3
    $\begingroup$ @user4119 I've tried to reword a little your post, and I corrected your $H_0$. It will be helpful if you can confirm that (a) parental response is the outcome, scored as a binary response (positive or negative "case"), (b) the scale you used is already a validated and unidimensional scale, that is working with a totalled score makes sense and reflect a single underlying construct (and all items were correctly reverse-scored, when needed). You said you cannot test for a gender effect because N's are even: it doesn't make sense at all. $\endgroup$ – chl Apr 11 '11 at 21:08
  • $\begingroup$ Thank you Jeromy, the wording make sense now. My instructor placed emphasis on only having 4 males and suggested I do a pooled variance. That is one of my concerns. The likert items are also assigned as ordinal. I am really tring to understand all of us this. $\endgroup$ – Donna Apr 12 '11 at 6:15
  • $\begingroup$ With having a binary response (positive or negative) does this mean to do a descriptives and frequencies analysis for a total sum etc.? Will I need to do any other tests that correlations? $\endgroup$ – Donna Apr 12 '11 at 6:25
  • $\begingroup$ I'm not sure what you mean when you say "want to assess whether social worker attitudes have an effect on parental responses." You don't mention collecting data on social workers' attitudes? How do get to the overall outcome of the case? $\endgroup$ – Andrew May 18 '11 at 22:47

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.