Tag Archive for: data

Meet Our Intern – Nadia

/in , /by

Nadia stands in front of the Sydney opera house at night with a dark blue sky My name is Nadia and I am excited to announce myself as the newest intern here at The Evaluator.

A bit about me: I’m currently studying Economics at the University of Liverpool, where I’ve developed an interest in Data analytics and statistics.

This internship presents an incredible opportunity for me to dive deeper into my interest of data analytics and apply the theoretical knowledge I’ve gained in a practical setting. What appeals to me about The Evaluator specifically is their core value of being creatively simple, I think the focus of anyone who handles data should be to ensure that any data can be easy enough for the average person to understand and useful enough that it can help them view their project or business in a way they may not have before.

Outside of studying my interests include music, film and travelling. In fact, I recently just spent 6 months living and travelling in Australia. I spent a semester studying in Melbourne and the rest of my time travelling across the east coast – I think my favourite memory from travelling abroad would be getting to go to the Melbourne Grand prix and meeting Daniel Riccardo as I’m a big fan of his. That or being chased by a possum was a fun memory also.

During my internship with The Evaluator my goal is to develop my data and report writing skills, both of which I feel I have already made a start on in the 2 months since I have been here. I’m excited to go forward and see what the new year has in store and I can’t wait to see what I can learn from Kirsty and the team.

 

Learning about data; Working with quantitative data

/in , /by

In a previous blog, we described quantitative data as data that can be counted or measured in numerical values. A spreadsheet is a good basis for organising data ready for analysis. Now let us imagine you have some data to hand, this is how you might want to try and use it.

Ways to analyse quantitative data.

Example 1: Looking at what the data tells you itself.

We generally start with calculating averages to consider the range of data (the highest score and the lowest score collected).  If we consider 2 sets of data (Data set A and Data set B):

From Data Set A, it is apparent that a similar experience was shared by the large majority of the audience because the average score is 40, but the range is from 36-44. That means the lowest answer was 36, which is quite close to 40 and the highest was 44 which is still quite close to 40.

However, in Data Set B, there was a much wider range of experience.  This should be questioned as to why experiences are so varied, illustrated by the wide range in the data.  Why did some individuals have such a higher score than others? Did some people have a different tutor, or venue, or have more time? There are many reasons why the range could be so high.

Example 2: Comparing data to other national data available.

Another possible way to compare data could be as in the worked example below where the Warwick Edinburgh Scale of Wellbeing was used to measure participant wellbeing.

Participants scores were collected over a period of time, firstly at the start of their engagement and then, at the end.  This graph represents a comparison in:

  • Each individual’s wellbeing from the start versus at the end. This is often a reflection of the impact of their engagement.
  • The individuals’ wellbeing who took part versus the national average
  • The individuals’ wellbeing score versus the NHS score of 40 which is believed to be indicative of poor wellbeing.
  • The individuals’ wellbeing versus one another.

Using quantitative data to help with forecasting & future decision making

Using percentages makes making comparisons easier to relate to and understand.

We use an example of a cinema and record audience attendance numbers of time.  Attendance could be affected by the popularity of the content of the film, seasonal trends or weather.  The trend line shows the overall audience is growing and then this line can be used to forecast on at the same trajectory.  This will help you to identify if you are likely to achieve your audience targets.

This information can also help you to make decisions about your capacity too and streamline your resources. For example, if you were thinking about moving a cinema location to a larger space, it would be worth looking at the trend line to think about venue size. If you were looking at two venues and one had a capacity of 100, and another 200, you might want to look at what your predicted audience size would be in six months’ time and negotiate the lease or hire accordingly. This is an example of making a data-driven-decision, something we are passionate about at The Evaluator.

Taking notice of negative space in data

It’s important to take note of the 70% who are agreeing but also to take note of the 30% who are disagreeing and find out why this is the case.  It is worth delving a bit deeper into on the minority and finding out what was the cause of their response.

At The Evaluator, we tend to represent 3 answers ‘yes’, ‘no’ and ‘prefer not to say’ in our reporting. If there’s a high number who indicate that they would ‘prefer not to say’ then it would be suggested they might insecure about completing the survey.  Often their indecision is explained in their qualitative answers and this is worth taking note of when creating future surveys.

Dealing with ‘satisficing’ survey responses.

This is the term we use when people respond with the answers that they think you are looking for and may not read all the questions.  An indication of satisficing is when respondents repeatedly choose 3 when presented with a 1-5 scale. We don’t come across these very often, as we spend a lot of time making sure our surveys are easy to complete, and varied, but if we do spot them, we will try and remove these answers from our data analysis.  It’s important to encourage honesty in answering survey questions.

A final tip

All data can be segmented but you do need to think about the time you have to spend on this as it is time consuming.  In segmentation ideally you are looking for what is more than 10% different to the average. In the example of the graph below, it is worth looking at segmenting to see the results demographics of the areas that are 10% above and 10% below the average line.

 

 

Learning about data; What is quantitative data?

/in , /by

What is Quantitative Data?

Quantitative data is data that can be counted or measured in numerical values.  As with qualitative data, there’s a good chance that you already have some collected for your organisation.

You might have collected some of the following:

  • Sign-in sheets
  • Feedback forms
  • Surveys
  • Polls
  • Social media statistics
  • Reports

We often find clients already have quite a bit of data they didn’t know they had collected!

The differences between primary and secondary quantitative data.

There is a distinction between primary and secondary quantitative data.  Primary data is the data that your organisation has collected directly, such as footfall counts or feedback forms. Secondary data is data someone else has collected, for example a national age profile, or a partner shares their footfall data. It can helpful for you to draw comparisons between your collected data and national averages to see how your organisation compares.

Overcoming the challenges of working with quantitative data

There are some challenges to working with quantitative data.  Often the biggest challenge is that it’s not collected in a format that makes it easy to compare to other collected data, or to secondary sources. The best solution for this is to plan in advance and use standardised questions at every opportunity to collect data. The answer format is also important so choosing a standard answer format will make it easier to compare data.

Tracking codes can be useful in identifying if you know the same person will be answering a survey multiple times so you can monitor their progress.  A tracking code can be created within a survey using data such as: a combination of a person’s date of birth and their initials.

It’s always important to date the data particularly for paper copies of surveys which makes identifying the event possible and the data relatable to that event. If one of your feedback forms reveals a problem with the venue or experience, you need to know the date on which that particular event happened to make sure you can address the problem. Don’t forget, feedback forms may be input or analysed as a batch of forms after a few months of collection so it may prove difficult to find out which venue the problem occurred at if you don’t have a way to check.

Managing personal data from surveys

GDPR (General Data Protection Regulation) are regulations which relate to how we retain and use personal data.  Within these regulations it is important to:

It’s important to maintain confidentiality and anonymity with personal information.  Recording date of birth and full name poses a risk to personal identity, however, recording only a date of birth is not identifiable. There are also additional regulations regarding collection data from under 16-year-olds. It is possible to collect identifiable information, but if you do so you need to ensure that the data is obtained with consent, is properly secured and then destroyed once no longer needed.

Thinking about when to collect data

Recording information in the moment is valuable so it can help to set up processes to ensure you don’t miss out! One tactic that works is to have a standard question you ask at the end of every event.  This, and the size of audience questioned can be collected for contextual purposes to see if the responses were representative of the larger audience.

Top tip – you don’t need to collect data from everyone!

Deciding how much data to collect

It’s important to consider whether sample sizes are large enough to provide you with sufficient data to base a decision on.  10% to 20% of the audience is usually a reliable sample size to base a decision on. If you hold many smaller events, it would be advisable to collect evidence from each event and consider it accumulatively to make decisions.

 

 

 

 

 

 

 

 

Learning about data; Working with qualitative data

/in /by

From time to time, we share a director’s blog post, where we share some learning about how to use your data (information) better. This is a post all about how to use qualitative data, that’s the information that is made up of words rather than numbers.

The chances are that your organisation already has a bank of this data but it may not be well organised and easily accessible. Below is a list of some of the many ways you might have collected qualitative data, possibly without realising it.

How credible is your qualitative data?

There is a distinction between primary and secondary data and it’s important to understand what you have collected as the credibility differs between the two.  Primary data is more credible and robust.  It is the data that people told you or have written themselves.  Secondary data is that which is overheard in discussion or something that someone has heard and told you about.

To check the credibility of your secondary data, you can apply the Rule of 3. If you’ve heard it from 3 different places or if it’s come from 3 different people independently then it can be considered as credible and representative of a thought or opinion.

How to make the qualitative data more usable? 

We use word or phrase frequency analysis to evaluate our qualitative data to look for patterns of frequency of words to identify common themes.  This can be done online using free tools, like this one,  and they just count the words for you.  We often use the top 10 words or phrases.  You can also use the ratio of positive to negative words that are used or how far down the feedback the first negative word appears.  Using these methods helps to quantify data and make it more digestible and can be used in marketing or to track changes over time.

Mind maps can be useful to illustrate and develop on the themes identified. You can just draw these freehand to have a look at what the main themes are.

Word clouds provide a visual representation at a glance of the qualitative data and this is also a resource that can be sourced freely online.  The most common words appear the biggest in the cloud, making this data easier for the reader to understand visually.

Top tip – try to be objective, it can be hard to hear negative comments but it is how we improve and know what to fix, and they’re often in the minority compared to positive comments.

How to use and share qualitative data?

This data can be used in many different ways:

We also use qualitative data to create case studies which illustrate people’s journeys and direct engagement.  Case studies can be shared at board level and to show case your project’s work for marketing purposes.  Case studies appeal to a wide audience and are particularly useful in attracting funders as voices are recorded and reflected in these studies.

In our experience as evaluators, a case study from a project that we evaluated got shared with funders (Green Recovery) who sent it on to The National Lottery Heritage Fund who then sent it on to the Department of Work and Pensions.  Case studies are impactful in that they record real voices and can attract publicity and raise awareness of change.

 

If all of this sounds like too much work and you don’t have time, get in touch. We are happy to have a chat about your individual requirements and to see how we could help.