Can I use Multiple Regression on my data?

#1
Hello!
I have an upcoming research project in which I have to examine the relationship between trait and state anxiety and the frequency a new mother checks her infant's development (FOC)
Can I use state and trait anxiety as predictor variables and FOC as an outcome in a multiple regression analysis? Or would there be a better suited statistical test to examine the link between state anxiety and trait anxiety on FOC?
I was also thinking it might be interesting to run a moderation to test if age moderates the r.ship between anxiety and FOC.
Any advice is greatly appreciated!
 

ondansetron

TS Contributor
#3
@hlsmith makes a good point. How is "frequency of new mother checks" defined and measured? Is it total checks in a 24-hour period? is it the number per hour? better explanation will be useful in making sure the model is specified appropriately since it sounds like a bounded count would be the dependent variable rather than an unbounded continuous variable.
 
#4
Thanks for your reply
@hlsmith makes a good point. How is "frequency of new mother checks" defined and measured? Is it total checks in a 24-hour period? is it the number per hour? better explanation will be useful in making sure the model is specified appropriately since it sounds like a bounded count would be the dependent variable rather than an unbounded continuous variable.
Thanks for your reply! I am very new to all of this so I appreciate the insight. Participants will be required to answer 3 questions regarding their FOC behaviour. The answers range from 0-3 on a likert type scale (never, rarely, sometimes, often etc) and the questions are along the line of " I worry my baby isn't developing as fast as other infants I see around me"
Would this work in meeting my assumptions for multiple regression? I'm a little confused as to what my variables are classed as. Is my outline variable continuous and are my predicator variables (anxiety) considered categorical or continuous (low, average, high) and if so, would I be better off using a 2-way ANOVA?
 
#5
Thanks for your reply

Thanks for your reply! I am very new to all of this so I appreciate the insight. Participants will be required to answer 3 questions regarding their FOC behaviour. The answers range from 0-3 on a likert type scale (never, rarely, sometimes, often etc) and the questions are along the line of " I worry my baby isn't developing as fast as other infants I see around me"
Would this work in meeting my assumptions for multiple regression? I'm a little confused as to what my variables are classed as. Is my outline variable continuous and are my predicator variables (anxiety) considered categorical or continuous (low, average, high) and if so, would I be better off using a 2-way ANOVA?
*outcome variable- not outline
 

ondansetron

TS Contributor
#6
Thanks for your reply

Thanks for your reply! I am very new to all of this so I appreciate the insight. Participants will be required to answer 3 questions regarding their FOC behaviour. The answers range from 0-3 on a likert type scale (never, rarely, sometimes, often etc) and the questions are along the line of " I worry my baby isn't developing as fast as other infants I see around me"
Would this work in meeting my assumptions for multiple regression? I'm a little confused as to what my variables are classed as. Is my outline variable continuous and are my predicator variables (anxiety) considered categorical or continuous (low, average, high) and if so, would I be better off using a 2-way ANOVA?
So FOC is frequency of checking, and is the dependent variable, I would say you could use an ordinal logistic regression (not a linear regression model). Both can be "multiple" in the sense that they can employ more than one x variable (predictor). For FOC, do you only have responses as "never", "rarely" and so on, or do you also have 0, 1, 2, 3, etc? Does a mother have only one value for her FOC or can the same person appear multiple times in the data?
 
#7
There are 3 questions related to FOC in ny survey (they measure development, feeding and crying) and participants would ideally give an indication on a likert type scale for each.
I was thinking to translate the responses into numbers 0= rarely- 3= often etc.... would this work for a multiple regression, or would an ordinal logistic regression be better suited?
Ultimately, I'm looking for the relationship between anxiety (state and trait) and FOC.
Thanks so much for your insight! I'm really confused about how to operationalise this!