登录查看更多内容

What's the right sample size and method for your stats project?

由人工智能和领英社区提供技术支持

Choosing the right sample size and method for your stats project can make a big difference in the quality and validity of your results. But how do you decide how many observations you need and how to collect them? In this article, we'll cover some basic principles and steps to help you plan your sampling strategy.

此文章中的业界达人

由社区从 4 条内容中精选。了解更多

Joel Ira Weller

Research Scientist at Agricultural Research Organization, The Volcani Center

1 Define your population and variables

Before you can determine your sample size, you need to define your target population and the variables you want to measure or compare. Your population is the group of interest that you want to generalize your findings to, such as customers, students, or voters. Your variables are the characteristics or outcomes that you want to analyze, such as satisfaction, performance, or preference. You should also specify the type of variable, such as categorical, numerical, or ordinal, as this will affect your sampling method and analysis.

添加您的观点

2 Choose your sampling method

There are many ways to select a sample from a population, but they can be broadly classified into two categories: probability and non-probability sampling. Probability sampling means that every element in the population has a known and non-zero chance of being included in the sample, such as random, systematic, or stratified sampling. Non-probability sampling means that the selection is based on some other criteria, such as convenience, quota, or judgment sampling. Probability sampling is generally preferred for statistical inference, as it allows you to estimate the sampling error and the confidence level of your results. Non-probability sampling may be used for exploratory or qualitative research, but it may introduce bias and limit the generalizability of your results.

添加您的观点

Bill Dudley

Professor Emeritus UNC Greensboro
举报内容
Although probability sampling is the gold standard, my experience over the past 30 years of quantitative research is the non-probability sampling is far more common. And research using Non-probability sampling often is employed for hypothesis testing.

已翻译

赞

3 Calculate your sample size

Once you have chosen your sampling method, you need to calculate how many observations you need to achieve your desired level of precision and power. Precision refers to how close your sample estimate is to the true population parameter, and it is usually measured by the margin of error or the confidence interval. Power refers to how likely you are to detect a meaningful difference or effect between groups or variables, and it is usually measured by the significance level or the p-value. To calculate your sample size, you need to specify these parameters, as well as the expected variability and effect size in your population. You can use online calculators, formulas, or software to perform the calculation, depending on the type of analysis you plan to conduct.

添加您的观点

4 Consider your resources and constraints

The final step in determining your sample size and method is to consider your practical limitations and ethical implications. You may have a limited budget, time, or access to your population, which may affect your sampling method and size. You may also have to deal with issues such as non-response, missing data, or outliers, which may reduce your effective sample size and introduce bias or error. Moreover, you should ensure that your sampling procedure follows the ethical principles of consent, confidentiality, and respect for your participants, and that you comply with any relevant regulations or guidelines.

添加您的观点

5 Review and adjust your sampling plan

After you have completed the previous steps, you should review and adjust your sampling plan as needed. You may find that your sample size is too large or too small for your objectives, or that your sampling method is not suitable for your population or variables. You may also encounter new challenges or opportunities that require you to modify or refine your sampling strategy. You should always document your sampling plan and rationale, and report any changes or limitations that may affect your results or interpretation.

添加您的观点

Sara Bonito

Associate @ J.P. Morgan | Applied Mathematics | Prev Deloitte, Microsoft
举报内容
Other things that may be useful to consider are the following: - Expected Response Rate: If your study involves surveys or questionnaires, estimate the expected response rate and adjust your initial sample size accordingly. - Effect Size: In hypothesis testing, determine the effect size you wish to detect. A larger effect size may require a smaller sample size to detect significant differences. - Potential Subgroups: If your analysis involves subgroup comparisons, ensure that your sample size is sufficient for each subgroup.

已翻译

赞

6 Here’s what else to consider

This is a space to share examples, stories, or insights that don’t fit into any of the previous sections. What else would you like to add?

添加您的观点

Joel Ira Weller

Research Scientist at Agricultural Research Organization, The Volcani Center
举报内容
Two major problems should also be considered: 1. The problem of multiple comparisions 2. Data mining With regards to the first, nearly all classical statistic theory is based on testing a single hypothesis. In modern studies, this is almost never the case, and in many studies hundreds or thousands of hypotheses are tested. Various methods have been proposed to deal with this problem. The second problem refers to the fact, that once an experiment is performed, the data may be used (or mis-used) to test additional hypothesis. There is a tendency to "play" with the data in order to obtain "statistical significance," which is generally consider the criterion for publication.

已翻译

赞
Anil Kumar Suthar

People Analytics Professional (Statistician)
举报内容
Scope of the underlying problem, considered hypothesis and objectives of study are vital determinants of sample size along with cost and resources available. It's always critical to determine the sample size in case of multiple hypotheses and wider scope. Also, complex studies require proper quota representation which could be achieved via effective study and questionnaires design.

已翻译

赞

Statistics

+ 关注

给文章评分

我们借助人工智能创建了此文章。您认为这篇文章怎么样？

很棒不太好

举报此文章

查看全部

What's the right sample size and method for your stats project?

1

2

3

4

5

6

1 Define your population and variables

2 Choose your sampling method

3 Calculate your sample size

4 Consider your resources and constraints

5 Review and adjust your sampling plan

6 Here’s what else to consider

Statistics

给文章评分

感谢您的反馈

更多Statistics相关文章

更多相关阅读内容

What's the right sample size and method for your stats project?

1

2

3

4

5

6

1 Define your population and variables

2 Choose your sampling method

3 Calculate your sample size

4 Consider your resources and constraints

5 Review and adjust your sampling plan

6 Here’s what else to consider

Statistics

给文章评分

感谢您的反馈

查看其他技能