Hi,
I came up with below hypotheses.
Null H1: there are no age difference in becoming A
Alternative H2: Younger people are more likely to become A
From what i have read so far putting all the numeric age data collected into certain categories (e.g., 18, 19, 21 into 18-25 category) seems to be the main way, however, I was wondering if there is a way I can run analysis (such as linear regression) without making those categories and instead putting them as they are (i.e., continuous, the age data range from 19 to 76) because I believe drawing the line (operationalization?) to create those categorical groups could be subjective. What do you guys think? Hope this question makes sense. Thank you. #spss #regression
"becoming A" is the response here. This is a binary variable.
"age" is continuous and is the predictor here.
A model for a binary response depending on a continuous predictor is called a (binary or binominal) logistic regression:
https://statistics.laerd.com/spss-tutorials/binomial-logistic-regression-using-spss-statistics.php
You will eventually get a coefficient for "age", what is a log odds ratio. This tells you, wow much the log odds of becoming A change per unit of age (per year, usually). If this coefficient is positive, the log odds increase with increasing age, that means: older people have a higer odds (probability) of becoming A than younger people. If the coefficient is negative, it's the other way around. You can test the coefficient to see if your data are sufficient to interpret the observed sign (beyond the observed data). If the test is significant, you can then conclude a positive (or negative, whatever the sign of the coefficient is) correlation between the probability of "becoming A" and age (just to note: the tested null hypothesis is that the log odds ratio is zero).
And: Yes !!! Never ever categorize a continuous variable! Never. A categorization can be done at the very, very end by the one who wants to practically apply your model for some purpose. But this is application, not research.
"becoming A" is the response here. This is a binary variable.
"age" is continuous and is the predictor here.
A model for a binary response depending on a continuous predictor is called a (binary or binominal) logistic regression:
https://statistics.laerd.com/spss-tutorials/binomial-logistic-regression-using-spss-statistics.php
You will eventually get a coefficient for "age", what is a log odds ratio. This tells you, wow much the log odds of becoming A change per unit of age (per year, usually). If this coefficient is positive, the log odds increase with increasing age, that means: older people have a higer odds (probability) of becoming A than younger people. If the coefficient is negative, it's the other way around. You can test the coefficient to see if your data are sufficient to interpret the observed sign (beyond the observed data). If the test is significant, you can then conclude a positive (or negative, whatever the sign of the coefficient is) correlation between the probability of "becoming A" and age (just to note: the tested null hypothesis is that the log odds ratio is zero).
And: Yes !!! Never ever categorize a continuous variable! Never. A categorization can be done at the very, very end by the one who wants to practically apply your model for some purpose. But this is application, not research.
How you enter it will depend on what you are looking at. For example, if A is a high school graduate, you would take into account some of the rules in the education system (so not many graduate before 16, but many do at 16-18, and some do as adults. Also, depending on your sample you might have different ages and some who will become A but haven't yet, and you may need to take this into account. So a 20 year old may graduate after you have stopped collecting data.
thank you Daniel Wright and Jochen Wilhelm for your responses! After reading your comments, I have come to realize there has been some misunderstanding on my side.
Jochen Wilhelm, I will conduct the logistic regression as this seems to be the type of analysis I was looking for thank you so much for this answer!
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Regression