Girl Guides

Girl Guides

by | Sujit Rathod -
Number of replies: 11

From the Guardian

FOR ALL STUDENTS

1. What is the outcome? Is this a prevalence or incidence figure?

2. What is the study design?

3. Calculate and interpret RRs for feeling safe in the North and the Midlands, using the South as the reference group.

4a. Calculate and interpret an RR for feeling safe for white girls, using girls of colour as the reference group.

4b. Calculate and interpret an RR for feeling safe for girls of colour, using white girls as the reference group.

FOR RETURNING STUDENTS

5. Comment on how sampling bias might have affected the 19% figure.

6. Comment on how measurement bias might have affected the 19% figure.

7. Comment on whether the RR figures are true and meaningful?

In reply to | Sujit Rathod

Re: Girl Guides

by | NICOL REDZO -
Thank you, Sujit for these interesting blogs.  Kindly find my attempt.

FOR ALL STUDENTS

1.    a)  What is the outcome?

Outcome:  any Self-reported safety in daily lives among girls aged 7 to 21 years.

1.    b) Is this a prevalence or incidence figure?

These are prevalence figures because the numerator includes all reported safety issues regardless of when they occurred and did not look at only recent safety reports so the figures are prevalence figures

 

2.    What is the study design?

The study surveyed women aged 7-21 years across the UK from March to April 2022 to estimate the prevalence of self-reported safety concerns. Data were collected at a single point in time (March to April 2022) therefore the design was cross-sectional.

 

3.    Calculate and interpret RRs for feeling safe in the North and the Midlands, using the South as the reference group.

·         The proportion in the North not feeling safe was 22%. Therefore those that feel safe were 100% - 22%  = 78% or 0.78

·         The proportion in Midlands not feeling safe was 19%. Therefore those that feel safe were  100% - 19% = 81% or 0.81

·         The proportion in the South not feeling safe was 16%. Therefore those that feel safe were 100% -16% =  84%  or 0.84

Risk ratio of feeling safe in the Northern area with the southern area as the reference

= risk safety in north / risk safety in south

=0.78/0.84

=0.9285

=0.93

Therefore the ratio that felt safe among girls and women aged 7-21 years in the Northern area of the UK was reduced by 7% compared to those in the southern area of the UK.

 

Risk ratio of feeling safe in the Midlands area with the southern area as the reference

= risk of safety in north / risk of safety in south

=0.81/0.84

=0.964

=0.96

Therefore the ratio that felt safe among girls and women aged 7-21 years in the Midlands area of the UK was reduced by 4% compared to those in the southern area of the UK.

 

4a. Calculate and interpret an RR for feeling safe for white girls, using girls of color as the reference group.

Given the risk of not feeling safe among white girls was 70%.  Therefore the risk of feeling safe is 100% -70%=30% or 0.30

The risk of not feeling safe among girls of color was 65%.  Therefore the risk of feeling safe is 100%-65%=35% or 0.35

 

RR for feeling safe for white girls

RR = risk of safety in white girls/ risk of safety in girls of color

=0.30/0.35

=0.857

=0.86

Among girls and young women aged 7-21 years, the risk of feeling safe was reduced by 14% among white girls and young women compared to girls of color.

 

 4b. Calculate and interpret an RR for feeling safe for girls of color, using white girls as the reference group.

 

RR for feeling safe for girls of color

RR= risk of safety in girls of color/risk of safety in white girls

=0.35/0.30

=1.166

=1.17

Among girls and young women aged 7-21 years, the risk of feeling safe was 17% more if are a girl or women of color than when you are a girl or a women of the white race.

FOR RETURNING STUDENTS

4.    Comment on how sampling bias might have affected the 19% figure.

If girls and young women participating in the study are different in terms of the outcome from those who were eligible and did not participate this could introduce sampling bias. This means that the 19% could be an underestimate if the girls and women across the UK that did not feel safe in their daily lives were less likely to participate in the study.

5.    Comment on how measurement bias might have affected the 19% figure.

Safety was self-reported. Given the sensitivity of the outcome, those that were more likely to experience abuse out of fear of the perpetrators would not report it or some may not want to be reminded of the abuses and may not respond to questions on safety issues. The 19% percent prevalence of girls not feeling safe in their daily lives could have been underestimated.

 

6.    Comment on whether the RR figures are true and meaningful?

It is difficult to tell if the RR figures are true and meaningful given that we are not given the target population size of girls and women aged 7 to 21 years across the UK and the actual number that participated (sample size).  If the sample that participated is not representative of the target population (small sample size) the estimates are likely to be inaccurate or biased. We needed confidence intervals to determine the level of precision.  



In reply to | Sujit Rathod

Re: Girl Guides

by | RANMINI SUMUDITA KULARATNE -
1. Outcome: feelings of safety and well-being among girls and young women aged 7-21 years across the UK
Prevalence figure

2. Study design: cross-sectional (survey)

3. RR for feeling safe in North c/t South = 22/16 = 1.38; RR for feeling safe in Midlands c/t South = 19/16 = 1.19. Prevalence of feeling safe 38% higher in the North and 19% higher in Midlands compared to the South of UK.

4. a) RR for feeling safe for white girls c/t girls of colour = 70/65 = 1.08. Prevalence of feeling safe among white girls approx 8% higher than among those of colour.
b) RR for feeling safe for girls of colour c/t white girls = 65/70 = 0.93. Prevalence of feeling safe among girls of colour approx 7% lower than among white girls.

5. Sampling bias - we do not know the sampling frame used; how participants were recruited. This affects generalisability (external validity) of findings. Participants may be of a relatively high socio-economic status and more likely to report feelings of safety and well-being, especially if more marginalised and disadvantaged populations were excluded from the survey. Figure may be lower than the true prevalence.

6. Measurement bias: self-reported feelings of safety and well-being subject to reporting (social desirability) bias by participants. Figure may be lower than true prevalence.

7. Cannot comment on this until we know more about the sampling methods used, the coverage of the survey, what questions were asked and whether these have been validated for estimation of safety and well-being.
In reply to | RANMINI SUMUDITA KULARATNE

Re: Girl Guides

by | Sujit Rathod -
Dear Nicol and Ranmini,

Lovely answers! I'm staring at the paragraph around 'feeling safe' and am wondering if the journalist has made an error in reporting? The 19% figure are those who "don't feel safe" and then the 22, 19, and 16% figures are those who "feel safe".

And having 22% in the North who "feel safe" makes them more likely to to feel safe than in the Midlands (19%) or South (16%).

Thoughts? -s

"Overall, almost one in five (19%) girls and young women who took part in the survey said they don’t feel safe at school, but those in the north were less likely to feel safe (22%), compared with 19% in the Midlands and 16% in London and the south. Girls and young women of colour are, however, less likely to feel safe at school than their white counterparts (65% against 70%)."
In reply to | Sujit Rathod

Re: Girl Guides

by | RANMINI SUMUDITA KULARATNE -
Hi Sujit

I think I misread the article.

1. Prevalence of feeling safe in North = 100-22 = 78%; in Midlands = 100 - 19 = 81%; in South 100-16 = 84%

Therefore RR for feeling safe in North c/t South = 78/84 = 0.93; RR for feeling safe in Midlands c/t South = 81/84 = 0.96. Prevalence of feeling safe 7% lower in the North and 4% lower in Midlands compared to the South of UK.

Would this be a correct interpretation?
In reply to | RANMINI SUMUDITA KULARATNE

Re: Girl Guides

by | Sujit Rathod -
Hi Ranmini - I think you're on the right track, but it's tough to be sure. As I noted, the journalist flips between figures about feeling safe and feeling unsafe.

This is a good example of why it is often a good idea to check the source material against what it reported in the media. And now you are developing the skills to be able to read these scientific papers.
In reply to | RANMINI SUMUDITA KULARATNE

Re: Girl Guides

by | MAR ESTUPINAN FERNANDEZ DE MESA -

Hello All, 

I found the RR question confusing, so it'd be helpful if someone can shed me a light. Thanks. 

By the way, these Epi in the news exercises are very helpful, thanks. 

Some ideas, here:

1. What is the outcome? Is this a prevalence or incidence figure?

Outcome: Feeling safe at school

Prevalence: almost one in five (19%) girl and young women

2. What is the study design?

Cross-sectional survey (Guirlguiding attitudes survey) March – April 2022, girls between 7 – 21 across the UK, n=3,000

3. Calculate and interpret RRs for feeling safe in the North and the Midlands, using the South as the reference group.

- The question is RR “for feeling safe”. 

- The article reports “[…] Overall, almost one in five (19%) girls and young women who took part in the survey said they don’t feel safe at school, but those in the north were less likely to feel safe (22%), compared with 19% in the Midlands and 16% in London and the south […]” 

I checked the study, and it reports “Around 1 in 5 (19%) girls and young women aged 11 to 21 don’t feel safe in school. This is higher in Wales (24%) and in the North of England (22%) than in other areas of the country. Girls of colour aged 11 to 16 are five percentage points less likely than girls who are White to feel safe at school (65% compared to 70%).”

Then in other chapter states “those in the North feel the least safe at school (22%) compared to 19% in the Midlands and 16% in London and the South.”

It feels the journalist has merged to sentences from to chapters of the report which makes the statement a bit confusing. 

If we take the journalist’s statement as of ‘not feeling safe’, then:

Proportion of GW feeling safe at school in the NW = 100% - 22% = 0.78

Proportion of GW feeling safe at school in the Midlands = 100% - 19% = 0.81

Proportion of GW feeling safe at school in London/South = 100% - 16% = 0.84

Therefore (?):

RR = Risk(exp) / Risk(unexp) 

North = 0.78 / 0.84 = 0.92

Midlands = 0.81 / 0.84 = 1.18 = 0.96

Compared to those in London and the South, 92% and 96% girls and young women in the North and Midlands feel safe at school, respectively

4a. Calculate and interpret an RR for feeling safe for white girls, using girls of colour as the reference group.

Proportion of GW of colour feeling safe at school = 65% =0.65

Proportion of White GW feeling safe at school = 70% = 0.7

RR = 70% (white) / 65% (women of colour) = 1.08

White girls and young women are slightly more likely (1.08 or 8%) to feel safe at schools than girls of colour.

4b. Calculate and interpret an RR for feeling safe for girls of colour, using white girls as the reference group.

RR = 65% (women of colour) / 70% (white) = 0.93

Compared to white girls and young women, 93% of women of colour feel safe at schools 

FOR RETURNING STUDENTS

5. Comment on how sampling bias might have affected the 19% figure.

Sampling variation – was there any over-represented subgroup (age, ethnicity, LGBTQI+)?

We don’t know the key characteristics of the participants (e.g., deprivation, geography (rural vs urban, what ethnicities were included and how where they grouped). Neither we know the characteristics of the non-respondents.

It’s also interesting to see that, although the sample covers 7-21 years old, the journalist reported a 22% value for those aged 11-16 in the second paragraph. However, it is unclear if the other figures reported refer to this age bracket or else. 

6. Comment on how measurement bias might have affected the 19% figure.

Reliability / Validity of the survey/ systematic errors

Respondent bias

7. Comment on whether the RR figures are true and meaningful

Unable to say much because 95% CI and p-values are not reported. Article neither cover confounders. 

One point that I think it’s worth reflecting is the way, how, and when journalists use the word ‘significant’ and whether they use it accurately (from stats point of view). E.g. they stated “The figure was significantly lower at 16% in London and the south.” Here they use ‘acute’ (what do they really mean by that?) “The problem was particularly acute among LGBTQ+ girls and young women, with almost two in five (37%) complaining about gender stereotyping at school.”


In reply to | MAR ESTUPINAN FERNANDEZ DE MESA

Re: Girl Guides

by | Sujit Rathod -
Dear Mar - welcome to Epi in the News, and thank you for this post!

I really appreciate how you checked the source material.

Also, in retrospect I should have asked about the Relative Risk (RR) for "feeling unsafe" instead of "feeling safe".

Your interpretations of the RR figure are good when the RR>1, but not quite right when RR<1. Check out what other students have written. -s
In reply to | Sujit Rathod

Re: Girl Guides

by | MAR ESTUPINAN FERNANDEZ DE MESA -

Thanks Sujit, very helpful. 

What I was thinking was e.g. 93% safe compared to a 100%, otherwise:

RR = Risk(exp) / Risk(unexp) 

North = 0.78 / 0.84 = 0.92

Midlands = 0.81 / 0.84 = 1.18 = 0.96

Compared to those in London and the South, the prevalence of feeling safe at school is 8% and 4% lower among  girls and young women in the North and Midlands, respectively.  Same for others with RR < 1. 

Would this be a more accurate reflection of the result?

Thanks. 

BW,

Mar

 

In reply to | Sujit Rathod

Re: Girl Guides

by | MANDO PHIRI -
1. What is the outcome? Is this a prevalence or incidence figure?
The outcome is feeling safe and happy in the school environment and represents the prevalence

2. What is the study design?

This was a cross section study among girls aged 7 to 21 years old

3. Calculate and interpret RRs for feeling safe in the North and the Midlands, using the South as the reference group.
North (RR) = 0.93

Girls and young women living in the North have a protective effect of 0.93 times more as compared those in the South.
Midlands (RR) = 0.96

Girls and young women living in the Midlands have a protective effect of 0.96 compared to those in the south.

4a. Calculate and interpret an RR for feeling safe for white girls, using girls of colour as the reference group.
RR (White girls) = 0.86.

Being a white girls and young woman has a protective effect on the feeling of being safe at school.

4b. Calculate and interpret an RR for feeling safe for girls of colour, using white girls as the reference group.
RR (Girls of Colour) = 1.17

Girl and young women of colour are 1.17 times more likely to fill unsafe in the school environment.
In reply to | Sujit Rathod

Re: Girl Guides

by | FATHIMA MINISHA -
Hi everybody...
this article does play around with the words too much... which is quite confusing...

1) Outcome- if girls feel safe at home or not (yes/no- binary outcome)
These are prevalence figures- its a crosssectional survey and looks at the outcome as present at the point of time...

2) Cross sectional survey

3) Feeling unsafe in North is 22%- so feeling safe is 78%
Feeling unsafe in Midlands is 19%, so feeling safe would be 81%
Feeling unsafe in south is 16% so feeling safe would be 84%
RR for feeling safe in north vs south= 0.93 (7% less feeling of safety in the north)
RR for feeling safe in Midlands vs south= 0.96 (4% less feeling of safety in the Midlands)
Of course we don't have further data to comment if this difference noted is a true finding or due to chance
Like you were commenting Sujit- the writer here is playing around with the words quite a bit... which i think would be really counterproductive...and fails to convey the message

4) Now for this question he uses figures for feeling safe- 65% and 70%. because the statement implies that girls of color have less % feeling safe compared to white.. so 65% should be for girls of color and so should be for feeling safe
RR for feeling safe white vs girls of color = 70/60= 1.16 (white girls feels safer by 16%)
RR for feeling safe girls of color vs white= 0.86 (girls of color 14% less likely to feel safe)

5) Really depends on the initial sampling frame used- the coverage and schools included (public vs private). Additionally for young girls- there is an issue of parents refusing consent... and the girls not included in the survey could be different from those in the study, hence the study sample not being representative of the general population of school girls in the UK..

6) the outcome is self reported- and safety is something quite difficult to define. the perceptions of the girls can be quite different- what one girls perceives as safe might be unsafe for another (unless they have an objective way to define the safety). There is also the issue of social desirability bias- or reported probably what the schools or parents wishes them to report. This would mean the 19% is actually an underreported figure.

7) Its difficult to say based on whats given in the article... Although anybody not well versed with how these things should be interpreted might feel the findings are of great impact.. We couldnt know unless we can have an idea about confounding, any bias (sampling, information, reporting), measurement errors, and unless chance is ruled out..

Fathima
In reply to | Sujit Rathod

Re: Girl Guides

by | JUDITH MARGARET BURCHARDT -
1. What is the outcome? Is this a prevalence or incidence figure?

Feeling safe at school. Point prevalence

2. What is the study design?

Cross sectional observational study

3. Calculate and interpret RRs for feeling safe in the North and the Midlands, using the South as the reference group.

(I wasn't sure I could do this as I didn't have data on feeling safe so I have produced figures for feeling unsafe)

0.22/0.16 = 1.38

Girls in the North are 38% more likely to feel unsafe at school than girls in London and the South

0.19/0.16 = 1.19

Girls in the Midlands are 19% more likely to feel unsafe at school than girls in London and the South


4a. Calculate and interpret an RR for feeling safe for white girls, using girls of colour as the reference group.

0.7/0.65 = 1.08

White girls are 8% more likely to feel safe at school than girls of colour.

4b. Calculate and interpret an RR for feeling safe for girls of colour, using white girls as the reference group.

0.65/0.7 = 0.93.

Girls of colour are 7% less likely to feel safe at school than white girls.

FOR RETURNING STUDENTS

5. Comment on how sampling bias might have affected the 19% figure.

Who answered this survey? Who was asked? I do not know. Were they all girl guides? It is unlikely that girl guides are representative of the population. The characteristics of girl guides may vary by region.

6. Comment on how measurement bias might have affected the 19% figure.

How was the question asked? Was it a leading question? Was it preceded by an explanation of the survey suggesting that girls might feel unsafe? Did the girls talk to each other before completing the questionnaire? What proportion of girls declined to answer the survey after having been asked?

7. Comment on whether the RR figures are true and meaningful?

We cannot know whether the RR figures are true and meaningful without knowing the answers to the questions about sampling and measurement bias above.
Accessibility
Reset All

Background Colour

Font Face

Font Size

1

Text Colour