I want to analyses the proportion of swimming sperm of three species of fish in two salinities. To analyse the proportion of swimming sperm in a Generalized Linear Model, I would use a Binary Logistic model with swimming sperm as dependent and the total number of sperm and the number of events occurring in each set of trials. However, we have tested sperm from each male in two different salinities i.e. a repeated design. Hence, I want to use a Generalized Linear Mixed Model with repeated measures. However, I fail to find how I do a Binary Logistic model with swimming sperm as dependent and the total number of sperm and the number of events occurring in each set of trials. Is it doable in SPSS? If so, how do I proceed? (Using the arcsine of the proportion of swimming sperm is not an option)
A binary logistic model is a special case of Binomial regression (aka Logistic), where the number of trials is 1. Your dependent variable is hence coded as 0 or 1.
If you have repeated trials and you know the number of trials, you can simplify the situation by summing the number of successes and failures (whatever that means in case of swimming fish sperm). In that case you'll be working with proportions. Should be doable in SPSS; fortunately I am an R user.
Martin Schmettow, thank you. I'm not absolutely sure if I understood your answer regarding "simplify the situation by summing the number of successes and failures".
The problem (for me) is that I know how to do it (without repeat) in the GzLM module, but I fail to find out how to do it in the GLzMM i.e. on repeated trials (freshwater and salt water). In a General Linear Model (GLM) it would be a repeated measures GLM with species as factor and salinity (fresh and salt) as repeat. Dependent variable would be the proportion of motile sperm.
My problem is not that I don't know what I want to do although I may fail to explain it (as well as use the correct terminology). My problem is that I can't find out how to do it in SPSS which is the program I am familiar with.
The data look like this.
Species A
N=x individuals
Sperm from each male was tested in freshwater as well as saltwater i.e. for each male, the sperm motility measurements was repeated in two salinitets.
Binary data: We know exactly how many sperm that was tested per individual and we know if a sperm was swimming or not i.e. each spermatozoa can be coded as 0 or 1 i.e of 3456 spermatozoa 2345 was motile.
Species 2.
Same as above
Species 3.
Same as above
Hello Ola Svensson. I think that what Martin Schmettow suggested is captured by the FIELDS sub-command for GENLINMIXED. Here is an excerpt from the help file (for SPSS 26):
TRIALS = NONE** | VALUE(number) | FIELD(field). If the model response can be expressed as the number of events occurring within a number of trials, then the TARGET keyword specifies the number of events and TRIALS specifies the number of trials. Use VALUE with the number of trials in parentheses if the number of trials is fixed for all subjects and FIELD with a field name in parentheses if the number of trials varies across subjects and there is a field containing the number of trials.
Source: https://www.ibm.com/support/knowledgecenter/en/SSLVMB_26.0.0/statistics_reference_project_ddita/spss/advanced/syn_genlinmixed_fields.html
HTH.
- Badges
- Science topic
- Similar topics
- Biological Science
- Zoology
- Ichthyology
- Vertebrates
- Fish