Why is sample size important in research?

Sample size is important because it affects the accuracy, reliability, and generalizability of
research findings.

1. Increases Accuracy

A larger sample is more likely to represent the population accurately and reduces the effect of
random chance.

Example: If you want to know the level of hopelessness among university students, asking 10
students may not reflect the true situation. Asking 300 students gives a more accurate picture.

2. Improves Reliability

With a larger sample, results are more stable and consistent. If the study is repeated, similar
results are more likely to be obtained.

3. Increases Statistical Power

A larger sample makes it easier to detect real differences or relationships.

Example: If seminary students and university students truly differ in suicidal ideation, a larger
sample helps you detect that difference.

4. Reduces Sampling Error

Sampling error is the difference between the sample result and the actual population value.
Larger samples generally have smaller sampling errors.

5. Enhances Generalizability

When the sample is large and representative, findings can be more confidently applied to the
broader population.

Problems with a Small Sample

 Results may be inaccurate.

 Important differences may not be detected.
 Findings may not represent the population.
 Statistical tests become less powerful.
One-Line Exam Answer

"Sample size is important because it determines the accuracy, reliability, statistical power,
and generalizability of research findings."

Issues with a Small Sample Size

1. Low Statistical Power

o Real differences or relationships may go undetected.
o Increases the risk of a Type II Error (failing to find an effect that
actually exists).

2. Less Accurate Results

o Results can be heavily influenced by a few unusual participants.
o Estimates may not reflect the true population values.

3. Poor Generalizability
o Findings cannot be confidently applied to the larger population.

4. Higher Sampling Error

o The sample may differ substantially from the population simply
by chance.

5. Unstable Results
o Repeating the study may produce very different outcomes.

6. Wider Confidence Intervals

o Less precision in estimating population parameters.

7. Greater Risk of Bias

o Small samples are more likely to be unrepresentative of the
population.

Example

If you compare hopelessness between university and seminary students using only 10 students
per group, the results may not accurately represent either group and important differences may
be missed.

Short Exam Answer

A small sample size can reduce statistical power, increase sampling error, limit
generalizability, produce unstable results, and lead to inaccurate conclusions.
Issues with a Very Large Sample Size

While larger samples are generally better, an excessively large sample can also create problems:

1. Higher Cost
o More money is needed for data collection, travel, printing, and
data entry.

2. More Time-Consuming
o Collecting, managing, and analyzing data takes much longer.

3. Resource Intensive
o Requires more researchers, materials, and administrative effort.

4. Statistically Significant but Practically Unimportant Results

o With a very large sample, even tiny differences can become
statistically significant.
o These differences may have little or no real-world importance.

Example: A difference of 0.2 points in hopelessness scores may become significant with
5,000 participants, even though it has little practical meaning.

5. Data Management Difficulties

o Handling and cleaning very large datasets can be complex and
increase the chance of errors.

6. Participant Management Challenges

o Monitoring data quality and ensuring consistent procedures
becomes harder.

Short Exam Answer

A very large sample size can increase cost, time, and resource requirements, create data
management difficulties, and may produce statistically significant results that have little
practical significance.

How to Decide Sample Size: Simple Explanation with Examples

Many students ask, "How many participants should I take?"

The answer is: Don't choose a random number. Think about these 6 factors first.
1. Research Design
The type of study affects how many participants you need.

Simple Survey (Descriptive Research)

If you're just describing something (e.g., levels of hopelessness among students), you may need
fewer participants.

Example: You want to know the average hopelessness level among university students.

 A sample of 150–200 students may be enough.

Experimental Research

If you're comparing groups or testing an intervention, you usually need more participants.

Example: You compare:

 Group A receives counseling

 Group B receives no counseling

You need enough participants in both groups to detect differences.

Key Idea:

More complex research = Larger sample needed.

2. Expected Attrition (Dropouts)

Some participants may:

 Refuse to complete the questionnaire

 Leave questions blank
 Drop out during the study

Therefore, collect more participants than you actually need.

Example:

You need 100 completed questionnaires.

You expect about 10% may be unusable.

Calculation:

 10% of 100 = 10
 Therefore collect 110 participants

If 10 drop out, you still have 100.

Key Idea:

Always plan for some loss of participants.

3. Statistical Test
Different analyses require different sample sizes.

Example 1: Correlation

You want to see whether hopelessness is related to suicidal ideation.

A moderate sample may be enough.

Example 2: Multiple Regression

You want to predict suicidal ideation using:

 Hopelessness
 Age
 Gender
 Social support
 Academic stress

Now there are many predictors.

You need a larger sample.

Example 3: ANOVA

You compare:

 University students
  Seminary students
 College students

More groups require more participants.

Key Idea:

The more advanced the analysis, the larger the sample should be.

4. Accessibility of Participants
Sometimes the ideal sample is impossible to obtain.

Example:

You want 500 seminary students.

But only 250 students attend the seminary and only 180 agree to participate.

You cannot magically get 500 participants.

Your sample must be realistic.

Example:

You are studying students in your university.

You can easily access:

 Psychology students
 Education students

But not students from another city.

Key Idea:

Choose a sample size that you can actually reach.

5. Number of Variables or Groups
The more things you compare, the more participants you need.

Example 1: Simple Study

Variables:

 Hopelessness
 Suicidal ideation

Only 2 variables.

A moderate sample may work.

Example 2: Complex Study

Variables:

 Hopelessness
 Suicidal ideation
 Depression
 Anxiety
 Stress
 Self-esteem
 Social support

Now the study is much more complex.

More participants are needed to get reliable results.

Example 3: Comparing Groups

Comparing:

 University students
 Seminary students

Need fewer participants than comparing:

 University students
 Seminary students
 College students
 Working adults
More groups = Larger sample.

Key Idea:

More variables and more groups require more participants.

6. Time and Resources

Sometimes a perfect sample is not practical.

Example:

Your supervisor says:

 600 participants would be ideal.

But you only have:

 1 month
 Limited travel budget
 No research assistants

Collecting 600 responses may be impossible.

A sample of 250–300 may be more realistic.

Consider:

 Time available
 Money available
 Printing costs
 Travel costs
 Data entry workload

Key Idea:

A smaller sample that you can actually complete is better than a huge sample that you
cannot finish.
Example from Your Research
Topic: A Comparative Study of Hopelessness and Suicidal Ideation among University Students
and Seminary Students

Before deciding sample size, ask:

1. Is it comparative research? → Yes

2. How many groups? → 2 groups
3. What statistical test? → Probably t-test or ANOVA
4. Can I access enough seminary students? → Check availability
5. Will some participants refuse? → Probably yes
6. Do I have enough time and resources? → Consider practical limits

A researcher might decide:

 150 university students

 150 seminary students

Total = 300 participants

This provides a balanced comparison while remaining manageable.

Easy Exam Answer

Sample size should be determined by considering the research design, expected attrition,
statistical test, accessibility of participants, number of variables or groups, and available
time and resources. A good sample size is both statistically appropriate and practically
feasible.

The Three Types of Sample Sizes: Simple Explanation in

Academic Language
When planning a research study, researchers often think about three different sample sizes: the
ideal sample size, the required (minimum) sample size, and the feasible sample size.
Understanding these concepts helps researchers choose a sample that is both scientifically sound
and practically achievable.
1. Ideal Sample Size
What is it?

The ideal sample size is the number of participants a researcher would like to have if there were
no limitations such as time, money, access, or resources.

It represents the best possible situation because larger samples generally provide more accurate
and representative results.

Example

Suppose you are studying hopelessness and suicidal ideation among university students in
Pakistan.

Ideally, you might want to survey:

 5,000 university students

 From different provinces
 From public and private universities

This large sample would provide a highly accurate picture of the population.

Why is it called "Ideal"?

Because it is what the researcher wants, not necessarily what they can obtain.

Key Idea

Ideal sample size = The dream sample size.

2. Required (Minimum) Sample Size

What is it?

The required sample size is the smallest number of participants needed for the study to produce
statistically reliable results.

Researchers determine this using:

 Statistical formulas
  Power analysis
 Previous research
 Research design

If the sample is smaller than the required size, the study may fail to detect real differences or
relationships.

Example

Suppose you want to compare:

 University students
 Seminary students

Using an independent samples t-test.

A power analysis indicates that you need at least:

 64 participants
 32 in each group

This means 64 is your required sample size.

Why is it important?

Without enough participants:

 Results become unreliable

 Statistical tests lose power
 Important findings may be missed

Key Idea

Required sample size = The minimum number needed for valid statistics.

3. Feasible Sample Size

What is it?

The feasible sample size is the number of participants that the researcher can realistically
collect.
It depends on practical limitations such as:

 Time
 Budget
 Accessibility
 Research assistants
 Population availability

Example

You need 64 participants according to power analysis.

However:

 You only have one month

 You can only access one university and one seminary

After considering these limitations, you determine that you can realistically collect data from:

 80 participants
 40 university students
 40 seminary students

This becomes your feasible sample size.

Key Idea

Feasible sample size = The realistic sample size.

Relationship Between the Three

Imagine conducting a study on hopelessness among students.

Sample
Type
Size

1,000
Ideal
students

Require
120 students
d
 Sample
Type
Size

Feasible 150 students

In this case:

 You want 1,000 students (ideal)

 You need at least 120 students (required)
 You can realistically collect data from 150 students (feasible)

This is a good situation because the feasible sample exceeds the required sample.

Main Goal
Researchers should try to make their feasible sample size equal to or larger than the required
sample size.

Good Situation

Required = 100

Feasible = 120

✓ Study can proceed.

Bad Situation

Required = 100

Feasible = 50

✗ Statistical power may be insufficient.

How Should Researchers Justify Their Sample
Size?
A researcher should never say:

"I chose 100 participants because it seemed like a good number."

This is not scientifically acceptable.

Instead, researchers should provide a logical explanation.

1. Based on Previous Studies

Researchers often examine similar studies and use comparable sample sizes.

Example

A previous study on hopelessness used 200 participants.

You may justify:

"The sample size was selected based on previous studies investigating hopelessness among
university students."

Why is it acceptable?

Because similar research has already shown that the sample size is appropriate.

2. Based on Accessible Population

Sometimes the total population is limited.

Example

You are studying seminary students.

Only 250 students are enrolled in the seminary.

You may decide to survey all 250 students.

Justification

"The sample size was determined based on the accessible population available to the researcher."

3. Based on Power Analysis

This is one of the strongest methods.

Power analysis calculates the minimum sample required for statistical testing.

Example

A power analysis indicates that:

 α = .05
 Power = .80
 Medium effect size

Required sample = 64 participants

Justification

"Sample size was determined through power analysis, which indicated a minimum requirement
of 64 participants."

4. Based on Number of Groups or Variables

More groups and variables require larger samples.

Example

Study 1:

Compare:

 University students
 Seminary students

Only 2 groups.
Study 2:

Compare:

 University students
 Seminary students
 College students
 Working adults

Now there are 4 groups.

A larger sample is needed.

Justification

"The sample size was increased to ensure adequate representation across multiple groups."

5. Based on Expected Attrition

Researchers often expect some participants to:

 Withdraw
 Leave questionnaires incomplete
 Provide unusable data

Therefore, extra participants are recruited.

Example

Required sample = 100

Expected attrition = 20%

Calculation:

100 + 20 = 120

Justification

"The target sample size was increased by 20% to account for expected attrition."
Weak Justifications to Avoid
Poor Example 1

"I selected 100 participants because it sounded sufficient."

No scientific basis.

Poor Example 2

"My friend used 150 participants."

Different studies require different sample sizes.

Poor Example 3

"This number was easy to manage."

Convenience alone is not a valid justification.

Example from Your Research Topic

Research Title: A Comparative Study of Hopelessness and Suicidal Ideation among University
Students and Seminary Students

Ideal Sample

 1,000 students from multiple institutions

Required Sample

 200 participants based on power analysis

Feasible Sample

 250 participants available from accessible institutions

Good Justification

"A power analysis indicated a minimum sample size of 200 participants. To account for possible
attrition and improve statistical power, the target sample was increased to 250 participants. This
sample size was feasible given the accessible population available to the researcher."

Easy Exam Summary

Concept Meaning

Ideal Sample The sample size the researcher would like to have under
Size perfect conditions.

Required
The minimum number needed for valid statistical analysis.
Sample Size

Feasible Sample The number that can realistically be obtained given

Size practical limitations.

Good Based on power analysis, previous studies, accessible

Justification population, attrition, or number of groups.

Random numbers, personal preference, or copying another

Bad Justification
study without rationale.

Golden Rule: A good sample size should be scientifically adequate (required) and practically
achievable (feasible), while aiming as closely as possible toward the ideal sample size.

Here is a simple, clear, and more logical explanation of your three topics. I’ll keep the
psychology meaning intact but make it easier to understand why sample size changes in each
case.

🟦 Topic A: Depression among undergraduate

students in Lahore
💡 Simple idea

You are trying to estimate how common depression is in all university students in Lahore.
🧠 Why sample size must be fairly large?

Think of Lahore universities like a big mixed soup:

 public + private universities

 rich + middle + low-income students
 different stress levels, lifestyles, cities, departments

If you only take a small spoonful of soup, you might miss important ingredients.

📌 Logical reasons for larger sample:

1. More diversity = more students needed Different universities and students behave
differently. A small sample cannot represent all groups fairly.

2. You are estimating “how many have depression” To estimate a percentage (like 20%), you
need enough people so the result is stable and not random.

 Small sample → 10–20 people can completely change the percentage

 Large sample → results become more stable and trustworthy

3. If you compare groups (male vs female, public vs private) Each group must have enough
people.

👉 Example:

 If only 20 females and 20 males → comparison is weak

 If 100 females and 100 males → comparison is meaningful

4. Not everyone will respond Some students refuse or skip sensitive questions.

So researchers collect extra people in advance to make sure final data is still enough.

✅ Simple conclusion:

You need a few hundred students because:

Depression rates must be measured accurately across many different types of students, not just a
small group.

🟩 Topic B: Social media addiction vs self-

esteem
💡 Simple idea

You are checking:

“Does more social media use mean lower self-esteem?”

This is a relationship (correlation) study.

🧠 Why sample size matters a LOT here?

Because correlation is very sensitive to mistakes.

📌 Logical reasons:

1. Small samples give fake relationships If you only study 20 students:

 one “extreme” student can create a false strong link

 or hide a real one

👉 So results become unreliable.

2. Real effects are usually small In psychology:

 social media vs self-esteem is usually weak (small effect)

So you need more people to “see” the pattern clearly.

3. Individual differences are huge Students differ in:

 personality
 family background
 phone habits
 confidence levels

More participants help balance these differences.

4. Measurement is not perfect Questionnaires are not 100% accurate. More participants =
errors cancel out = clearer pattern.

5. You may want to control other variables Like gender, age, or stress. More data helps
separate their effects.

✅ Simple conclusion:

You need around 200–300 students because:

Correlation is unstable in small samples and needs many participants to show the true
relationship clearly.

🟨 Topic C: Mindfulness training and stress

reduction
💡 Simple idea

You are testing:

“Does mindfulness reduce stress?”

This is an experiment, so you compare groups.

🧠 Why sample size depends on design?

Here the key issue is: 👉 How many groups you have and how people are assigned

📌 Logical explanation:

1. One group (before vs after)

You measure:

 stress before mindfulness

 stress after mindfulness

👉 Same people = easier comparison

👉 Less noise = smaller sample needed

✔ Because each person acts as their own control.

2. Two groups (mindfulness vs control)

Now you compare:

 Group A: mindfulness
 Group B: no mindfulness

👉 Different people = more variation

👉 Harder to detect difference

So you need more participants

3. Why two-group study needs more people?

Because:

 people naturally differ (stress, personality)

 groups are not identical
 differences can hide the real effect

So you need more data to “balance” randomness.

4. Dropout problem
In real life:

 some students stop attending mindfulness sessions

 others miss follow-ups

So you must start with extra participants.

5. More groups = more sample needed

If you add:

 mindfulness group
 placebo group
 control group

👉 comparisons increase → sample must increase

✅ Simple conclusion:

 One group (pre-post): fewer students needed

 Two or more groups: more students needed

Because:

Between-group differences are harder to detect than within-person changes.

⭐ Final Easy Summary

 Depression study (A): needs large sample because population is
diverse and you want accurate prevalence
 Correlation study (B): needs medium-large sample because
relationships are weak and unstable
 Experiment study (C): sample depends on number of groups and
design complexity
If you want, I can also turn this into: ✔
✔
✔ or

A correlation study needs more than a small sample because correlation is very sensitive to
random variation and misleading patterns when data is limited.

Here is a simple logical explanation:

📌 Why correlation needs a larger sample size

1. Small samples give unstable results
In a small sample (e.g., 15–30 people), one or two unusual participants can completely change
the correlation.

Example:

 If one student uses social media 10 hours/day and has very low self-
esteem
→ it can create a strong negative correlation even if the real
relationship is weak
 Or it can cancel out a real effect

👉 So the result is not reliable or stable

2. Correlation depends on patterns, not single scores

Correlation is about a trend across many people, not individual cases.

 Small sample → pattern is unclear

 Large sample → pattern becomes visible and consistent

👉 You need enough data points to “see the line” clearly.

3. Real-life correlations are usually weak
In psychology and social sciences:

 most relationships are small (e.g., r = 0.10 to 0.30)

Small effects are hard to detect in small samples.

👉 If sample is too small:

 you may wrongly conclude “no relationship exists”

4. Random error is higher in small samples

Every measurement has error (questionnaires, self-reports).

 Small sample → errors dominate results

 Large sample → errors balance out

👉 So larger samples give a more accurate estimate of the true relationship.

5. Restricted range problem

Small samples often come from one group (e.g., one class or school).

This causes:

 similar levels of social media use

 similar self-esteem

👉 Less variation = weaker or misleading correlation

6. Outliers distort results easily

In small samples:

 one extreme person can create or destroy a correlation

In large samples:
  outliers have less impact

⭐ Simple conclusion
Correlation needs a larger sample because:

It is a pattern-based analysis, and small samples are too unstable, too affected by random error,
and too easily distorted by outliers to show the true relationship.

If you want, I can also give you: ✔

✔
✔ or

For a study like “Prevalence of depression among university students in Lahore”, the sample
size is not chosen randomly. It depends on several logical and statistical factors that determine
how accurate and reliable your results will be.

📌 Factors affecting sample size

1. 🎯 Study objective (prevalence estimation)
Since you are estimating how common depression is, you need a sample large enough to:

 get a stable percentage (e.g., 20% depressed)

 reduce random fluctuation

👉 If the sample is small, the prevalence may change too easily (e.g., 10% vs 30%)

2. 🌍 Population size and diversity

Lahore has:

 many universities (public + private)

 different departments
 different socioeconomic backgrounds
  male/female variation

👉 More diversity = larger sample needed

Because you must represent all types of students fairly.

3. 📊 Expected prevalence rate

If previous research suggests:

 depression rate is ~15–30%

Then:

 moderate sample is needed (often 300–600+)

👉 If prevalence is rare or unknown → you need a larger sample to detect it accurately.

4. 📉 Desired precision (margin of error)

This is how “accurate” you want your estimate.

Example:

 ±5% error → larger sample needed

 ±10% error → smaller sample may be acceptable

👉 More precision = bigger sample

5. 🔢 Confidence level (usually 95%)

Most studies use:

 95% confidence level

👉 Higher confidence (99%) → requires larger sample

Because you want to be more sure about results.
6. 🧪 Sampling technique
Different methods affect sample size:

 Simple random sampling → smaller sample needed

 Cluster sampling (by university/classes) → larger sample needed
(because students in same class are similar)

7. 📚 Subgroup analysis
If you want to compare:

 male vs female
 public vs private universities
 different age groups

👉 Each subgroup must have enough participants (usually 50–100+ each)

So total sample increases.

8. ❌ Non-response and missing data

In mental health surveys:

 some students refuse

 some skip questions
 stigma may reduce response rate

👉 So researchers oversample by 20–30%

9. ⚖️Measurement tool reliability

Depression scales (like PHQ-9) are good but not perfect.

👉 More sample = reduces measurement error impact

Smaller sample = unstable results
⭐ Simple conclusion
Sample size for depression prevalence in Lahore universities depends on:

population diversity, expected prevalence, desired accuracy, sampling method, subgroup

comparisons, and non-response rate.

If you want, I can also: ✔ give you a perfect exam answer (5 or 10 marks)
✔ or calculate an approximate sample size using formula
✔ or write it in APA research proposal style

When you compare groups in a study, it directly increases the required sample size because
comparisons need enough data in each group to be fair and statistically reliable.

📌 How “comparing groups” affects sample

size
1. ⚖️Each group must be big enough on its own
When you compare (e.g., males vs females, or treatment vs control):

 You are not analyzing one group anymore

 You are analyzing two or more separate samples

👉 So instead of one sample size, you now need:

 Group A sample size + Group B sample size

If either group is small → comparison becomes weak.

2. 📉 More variability between groups

Different groups naturally differ in:

 personality
 background
 stress levels
  environment

👉 This increases “noise” in data

So you need more participants to clearly see real differences

3. 🎯 Detecting differences requires power

In comparison studies, you are testing:

“Is Group A significantly different from Group B?”

To detect this difference reliably:

 small sample → weak statistical power

 large sample → strong ability to detect real differences

👉 So sample size must increase to avoid false “no difference” results.

4. 📊 Smaller groups = unstable means

If a group has few people:

 average score (mean) becomes unstable

 one extreme person can change results

Example:

 10 students → one depressed student changes average a lot

 100 students → one student has little effect

5. ⚖️Equal group size improves accuracy

Comparisons work best when groups are balanced:

 50 vs 50 is better than 90 vs 10

If groups are unequal:

 statistical error increases

  you often need extra total sample to compensate

6. 🔍 More groups = more sample needed

If study has:

 2 groups → moderate sample needed

 3+ groups → even larger sample needed

Because each additional comparison increases complexity and error risk.

⭐ Simple conclusion
Comparing groups increases sample size because:

You need enough participants in each group to reduce random variation and reliably detect real
differences between groups.

If you want, I can also give you: ✔ a 2-mark short answer

✔ a 5-mark exam answer
✔ or a diagram-style explanation for revision 👍