What counts as a selection bias in this situation?

[happypolarbear45]

TL;DR

Does this count as a selection bias, or is it not a issue?

There have been many studies done to find out if cell phones cause infertility. Most have found that there is a positive connection between phone use and infertility. However, most studies are done at fertility clinics and their study population is fertility patients. My question is, does the fact that the study population is fertility patients create a selection bias? Or, is the study still valid as even though the population is already experiencing fertility problems, it is still a level playing field as people who use phones the most are 'more infertile' than others?

I am not very farmiliar with what would count as a bias and would appritaiate some input and advice on what others think about the possible selection bias.

Thanks for reading this!

 

[BillTre]

There have been many studies done to find out if cell phones cause infertility.

This is not convincing to me.
Your should provide a reference to some real peer reviewed, published article.
It is the basis for your whole post and looks wrong to me.

 

[Ygggdrasil]

Summary:: Does this count as a selection bias, or is it not a issue?

There have been many studies done to find out if cell phones cause infertility. Most have found that there is a positive connection between phone use and infertility. However, most studies are done at fertility clinics and their study population is fertility patients. My question is, does the fact that the study population is fertility patients create a selection bias? Or, is the study still valid as even though the population is already experiencing fertility problems, it is still a level playing field as people who use phones the most are 'more infertile' than others?

I am not very farmiliar with what would count as a bias and would appritaiate some input and advice on what others think about the possible selection bias.

Thanks for reading this!

It's hard to comment on such studies without seeing the actual study, but I could think of several biases that could lead to a false conclusion that cell phones cause infertility. Presumably, the observation of such studies would be that patients at infertility clinics use cell phones more than the general population. However, there are many differences between people who go to infertility clinics versus the general population. For example, people in infertility clinics would tend to be more likely to be married, be of a specific age range (likely on the older end of childbearing years, so late thirties), and probably richer (e.g. in a financial position to raise a family, having access to insurance that covers fertility treatments [or being rich enough to foot the bill for fertility treatments not covered by insurance], and/or in certain professional careers that would cause the patients to delay childbearing to later years) than the general population. Any of these factors could lead the population of patients at infertility clinics to have a higher cell phone usage than the general public without cell phones having any effect on fertility.

 

[happypolarbear45]

It's hard to comment on such studies without seeing the actual study, but I could think of several biases that could lead to a false conclusion that cell phones cause infertility. Presumably, the observation of such studies would be that patients at infertility clinics use cell phones more than the general population. However, there are many differences between people who go to infertility clinics versus the general population. For example, people in infertility clinics would tend to be more likely to be married, be of a specific age range (likely on the older end of childbearing years, so late thirties), and probably richer (e.g. in a financial position to raise a family, having access to insurance that covers fertility treatments [or being rich enough to foot the bill for fertility treatments not covered by insurance], and/or in certain professional careers that would cause the patients to delay childbearing to later years) than the general population. Any of these factors could lead the population of patients at infertility clinics to have a higher cell phone usage than the general public without cell phones having any effect on fertility.

Thanks for the reply! So, would you say that because of this it isn't sensible to make any conclusions out of a study that uses fertility clinics as it's population?

Also, studies like https://www.researchgate.net/publication/6184009_Evaluation_of_the_effect_of_using_mobile_phones_on_male_fertility are ones like that. They monitored 304 men from 2004-2006, divided them into groups who don't use phones and groups that do, and found that cell phones negitavley affect sperm count. However, they used fertility clinics as their population and their surveys don't ask for much detail:

The survey consisted of questions concerning the place of residence (division: rural area, town with a population of up 50,000; a large city with a population of over 50,000), age, smoking habits, occupation and use of phone.

This seems to ignore certian factors like diet, exersise, bathing habits, position of phone, etc. What do you think?

 

[Ygggdrasil]

Thanks for the reply! So, would you say that because of this it isn't sensible to make any conclusions out of a study that uses fertility clinics as it's population?

Also, studies like https://www.researchgate.net/publication/6184009_Evaluation_of_the_effect_of_using_mobile_phones_on_male_fertility are ones like that. They monitored 304 men from 2004-2006, divided them into groups who don't use phones and groups that do, and found that cell phones negitavley affect sperm count. However, they used fertility clinics as their population and their surveys don't ask for much detail:

The survey consisted of questions concerning the place of residence (division: rural area, town with a population of up 50,000; a large city with a population of over 50,000), age, smoking habits, occupation and use of phone.

This seems to ignore certian factors like diet, exersise, bathing habits, position of phone, etc. What do you think?

Yes, the study you cite is an observational study in which the researchers take a group of people, categorize the group of people into different groups, and analyze differences between the two groups. While these studies can be useful, they are also subject to confounding variables that can make it difficult to make definitive conclusions between the studies. For example, there are probably other differences between the group of people who use cell phones and the group of people who don't (e.g. in diet and exercise), and it could be that those differences are behind the difference in sperm count and not cell phones.

Stronger evidence would come from randomized controlled trials where people in a population are randomized between two conditions (e.g. using cell phones vs not using cell phones) and monitored over time. Randomization of the population would help eliminate any underlying differences between the two populations, so that the effect of the experimental variable can more clearly be observed.

For more information see: https://www.iwh.on.ca/what-researchers-mean-by/observational-vs-experimental-studies

 

[happypolarbear45]

Thank you so, so much for the reply! Can I clarify something? You said 'Stronger evidence would come from randomized controlled trials where people in a population are randomized between two conditions (e.g. using cell phones vs not using cell phones)', how is that different to what the study I shared did, as they divided subjects into groups of phone users and non phone users? Thanks again so much for explaining this to me!

 

[Ygggdrasil]

Thank you so, so much for the reply! Can I clarify something? You said 'Stronger evidence would come from randomized controlled trials where people in a population are randomized between two conditions (e.g. using cell phones vs not using cell phones)', how is that different to what the study I shared did, as they divided subjects into groups of phone users and non phone users? Thanks again so much for explaining this to me!

The difference is randomization. In the study, subjects were grouped based on their own choices of whether or not they used cell phones. In a randomized controlled trial, subjects would be randomly assigned to use cell phones or not use cell phones.

 

[jim mcnamara]

The answer to your question: randomized control trial. Way more complex than you might think.

https://en.wikipedia.org/wiki/Randomized_controlled_trial -RCT

The idea is that you control every variable you can possibly think of that may affect the result - examples:
gender, age, genetics, predisposing medical conditions, country, maybe socioeconomic status, education ...etc., etc. These trials are expensive to design, implement, and execute. So it is not a casual thing, as in "Hey, I've got an idea, let us test it this week!"

The Coronavirus vaccines RCT's were and are an example: thousands of people, nobody knows who is getting what drug or placebo. Was every person in the study they seropositive at the start, and so on. So, many people apply to participate and the testers have to sort through mountain of information to get a qualified test population. And then carefully monitor the program as it runs.

If you have questions please consider reading the link. First.

 

[happypolarbear45]

The difference is randomization. In the study, subjects were grouped based on their own choices of whether or not they used cell phones. In a randomized controlled trial, subjects would be randomly assigned to use cell phones or not use cell phones.

Thank you so so much! So should I not make a conclusion out of this evidence due to its flaws?

 

[Fervent Freyja]

Thank you so so much! So should I not make a conclusion out of this evidence due to its flaws?

No, you shouldn’t!

 

[happypolarbear45]

The answer to your question: randomized control trial. Way more complex than you might think.

https://en.wikipedia.org/wiki/Randomized_controlled_trial -RCT

The idea is that you control every variable you can possibly think of that may affect the result - examples:
gender, age, genetics, predisposing medical conditions, country, maybe socioeconomic status, education ...etc., etc. These trials are expensive to design, implement, and execute. So it is not a casual thing, as in "Hey, I've got an idea, let us test it this week!"

The Coronavirus vaccines RCT's were and are an example: thousands of people, nobody knows who is getting what drug or placebo. Was every person in the study they seropositive at the start, and so on. So, many people apply to participate and the testers have to sort through mountain of information to get a qualified test population. And then carefully monitor the program as it runs.

If you have questions please consider reading the link. First.

Thank you so much! May I ask you a question? I read a study that monitored cell phone users from 2008-2011 and found that phones cause infertility. However, in their study they said:

The study population consisted of 286 men who attended the infertility clinic for diagnostic purposes and who had normal semen concentration of 20–300 M ml−1 or with slight oligozoospermia (semen concentration of 15–20 M ml−1)

Forgive me if this is the wrong assumption, but surely excluding those who have abnormal semen concentration creates a selection bias, as the study only looked at those who have healthy semen concentration, and surely if phones cause infertility then those affected would have abnormal semen concentration. Although they did find a correlation between phone use and infertility, could the fact that they excluded abnormal semen concentration be a selection bias?

 

[jim mcnamara]

I think what we are trying to tell you: unless there is good randomized trial, you can only guess about the true meaning of the study - if there is one. So look at the pyramid:

 

[jim mcnamara]

The interface to PF graphic images is not cooperating.
You can think of this as the hierarchy of quality of information you get from a given study. Every study you have asked about is case reports. The topmost level is used to make medical standards of care.

For example:
This is why nutrition news reports seem conflicting - because they are. They are mostly case studies.

So. It does not matter about selection bias, because there are too many other "loopholes" for the research you mentionto have a lot of import. This loophole thing is properly called 'confounding factor(s)'. This means you can't draw out what you want because way too many other things can claim they are just as important. Selection bias being one of many.

Ignore case studies like this unless you are writing term paper for school.

Since you are focused on cellphones at the moment, here is an RCT on the subject. See what they conclude about cellphones and users.

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4636032/