You’re conducting statistical research. How can you present your data in a way that’s easy to understand?

Consider your audience and canvas space. If your audience is well-versed in the subject, avoid excessive basic information. In a broader setting, like a public forum, provide ample context for a wider audience. Start with basic descriptive statistics, creating a foundation before delving into advanced models. Always balance the depth of information with the level of familiarity your audience has with the topic, ensuring clarity and engagement in your presentation of statistical research.

One of the first steps I like to do when presenting data is creating a story. During our research, we, as statisticians, often investigate different aspects and discover data correlations that might steer our investigations in unforeseen directions. This is the essence of research. However, when presenting data after the last research question is answered, the story we tell must be coherent, or else it is extremely difficult to follow for someone who sees the data for the first time.

Effective data presentation involves understanding your audience and the data's nature. Tailor your approach to audience familiarity and strike a balance between simplicity and detail. Use visuals like charts for trends, tables for details. Maintain consistency for coherence, ensure accessibility, and collect feedback for iterative improvement, enhancing clarity and effectiveness in your data presentations.

Apply the Covey principle of beginning with the end in mind. Your end product is the clear communication of the lessons / findings learnt through the research. Start by writing down on paper what these lessons are, using a separate piece of paper for each lesson, and then reverse engineer the explanations/stages/processes etc. required to arrive at this finding. Only after you can clearly trace the route to each lesson/finding, consider which media (charts, tables etc.) most strongly support the message that you wish to communicate, taking the audience into consideration where possible. When the audience is general or heterogeneous in nature, apply the twin considerations of parsimony and clarity in distilling the messages you communicate.

Present statistical research with clear visualizations like charts, focusing on key findings and using summary statistics for simplicity. Provide context with real-world examples and avoid jargon. Use tables for detailed breakdowns and infographics for engagement. Incorporate storytelling to guide the audience, and ensure clear titles and labels. Tailor the presentation to the audience's familiarity with statistics for accessibility, aiming for a concise and impactful delivery of key insights.

Employ descriptive statistics like means, medians, modes, ranges, and standard deviations to summarize and describe the basic features of your data. This provides a straightforward overview of your dataset and helps set the context for your findings.

Just as the name implies, descriptive statistics are used to describe your data. The descriptive statistics you use will depend on the type and properties of the variable(s) you’re working with. For quantitative variables, measures of central tendency (mean, median, and mode) are often a good start. They identify the central point within a variable using a single value. Additionally, measures of variability (e.g., standard deviation and range) are useful for quantifying the spread of the data. For qualitative variables, frequency distribution and cross-tabulations are helpful. They organize the number of occurrence of each category of the variable(s) into a table. It is essential to understand your data to know the appropriate measure.

The kind and distribution of the data should be taken into account when employing descriptive statistics since they have an impact on the measure of central tendency. The deviation or error in data is significantly affected by the outliers. The sample size is taken into account when estimating the population through study. being conscious of the descriptive statistics' underlying presumptions. Lastly, utilizing boxplots and histograms to display the patterns underlying the data

The descriptive statistics is for continuous data such as age, weight etc. For categorical or data with categories such as gender, occupation etc you use frequencies and percentages to describe a data or trends in the data.

Focus on key measures like mean, median, mode, and standard deviation to summarize central tendencies and variability. Utilize visual aids such as histograms, box plots, or bar charts to enhance understanding. Provide clear labels and concise explanations to accompany your data presentation.

In statistical research, clarity is key. Use inferential statistics to make your data speak. Techniques like t-tests, ANOVA, and chi-square tests help test hypotheses, estimate parameters, and assess relationships. Remember, the right test for your data type is crucial. Always report with key statistics - p-values, confidence intervals, correlation coefficients, or effect sizes. This approach not only clarifies your findings but also underpins their reliability and relevance.

When you conclude about the larger population from a sample, you are doing inferential statistics. Inferential statistics involves estimating the characteristics of a population from a sample and finding evidence for or against a claim about the characteristics. Inferential statistics may involve comparing groups, finding confidence intervals for a population parameter, or finding statistically significant relationships between variables. The technique you use will depend on the question(s) you want to answer and the distribution of your data. Popular methods include confidence intervals, t-tests, ANOVA, Chi-square tests, and regression analysis. Make sure your data meets the assumptions of your chosen method for a valid conclusion.

While the previous section described univariate analysis, you need to move beyond single variable analysis to two or more variable analysis. The two variable analysis is called bivariate analysis which measures a relationship between two variables. The inferential statistics tells how certain you are that such relationship exists in the population under study.

Cuando sea posible y los datos los permitan, simplificar el funcionamiento del mundo a través de la estadística inferencial permite tomar decisiones importantes sobre la población en estudio. Al buscar comparar, comprobar o estimar, pueden ser indicadores de un uso de esta clase.

Inferential statistics plays important role in decision making but requires a deep understanding of your data and variables . When data is categorical and you have two variables one can use fishers or chisquare tests for proportions. If one requires to compare means then we can opt for tests like ANOVA ,T.tests etc . However one should be careful with the numerical data , certain assumptions like normality need to be met and can be assessed through tests like Shapiro or Kolgomorov smirnov tests . Also assumptions like homogeinity of variance need to be checked before moving on . However to skip all the labours of testing assumptions one can use the power of bootstrapping , simulations and permutations to come up with robust conclusions .

However visualization should be easy to understand and should not be too complex or too fancy which is quite difficult to comprehend.

Visualising data as an effective way of looking and summarising your data . We can use boxplots to capture certain distribution of numeric data across certain categories . This also helps us to capture any anomalies within our data such as outliers and open doors to more questions that could be answered by t.test or Anova. We can also capture relationships using scatterplots and then move on to inferential stats because of these visuals.

Use charts, graphs, and other visual tools to make complex data more accessible. Select visualizations appropriate for your data and audience, such as bar charts for categorical data or line graphs for trends over time.

There’s no substitute for plotting or otherwise visualizing raw (and eventually processed/final) data to see what your datasets actually look like. Consult with experts or the literature on any patterns you observe. Follow up on anything you see that you don’t understand.

In my experience, the visualisations are a good way to understand the data, shape and distribution by the analysis and easy to present to the audience.

Aim to pass a message, always. One way to do this is by cutting clutter. By that I mean to exclude any statistical result that does not contribute to the overall narrative that is being presented. This requires considerable emotional discipline. Sometimes it is difficult to remove results or visualisations when we consider the effort it took to produce them in the first place. However, there is more to gain from clarity than needless volume.

Break down complex data into more straightforward, more digestible pieces. Avoid overwhelming the audience with too much information at a time. Focus on presenting one key point or finding at a time.

Al tener una investigación estadística, sugiere que se tiene gran cantidad de información, gran cantidad de datos, que son medibles y son explicables juntos y por separado, sin embargo, para que la explicación sea efectiva y fácil de entender, debe pasar por un filtro que permita extraer los datos prioritarios que permitirán demostrar y darle sentido al objetivo investigativo en cuestión.

To understand your data, you need to be familiar with the context, true meaning of variables, how they were collected, why are there some missing values, etc. Comparison of numbers that you take for granted is not enough. Use metadata for your datasets and describe what they represent. Knowing what is behind a number will encourage you to tell stories about your data. That is what people appreciate and understand.

Accompany your data with clear explanations. Provide context, interpret the results, and explain why they are significant. Avoid jargon and technical language, especially when presenting to non-specialists.

To be able to adequately and clearly explain your data, first of all you must understand it yourself implicitly. A good tip is to print out your tables and charts and write on the printouts what patterns leap out at you or which you consider important. Highlight these and then write a sentence of two on the printout explaining why they are important and how they contribute to the central findings of the research. Do this process for number of tables and you have the heart of a discussion there and then. Type out a rough draft of this "discussion" and then cross reference it with the list of central findings from the research. Identify gaps i.e. findings not linked to tables etc. and plug those gaps. Then polish the final discussion.

Explaining data is important to enhance the audience's understanding. It also reduces complexity and increases transparency. Without context, data can be misunderstood or misinterpreted, which can lead to the audience losing trust in the data and, consequently, the entire study.

It is crucial to focus not only on the analysis results but also on the audience, the questions asked, and the big picture of why the questions are asked. This helps to bring a unique twist to the presentation and make conclusions and recommendations valuable and actionable for the audience.

When using data for persuasion, it is important to tell a cohesive story and use visualizations to demonstrate how the data support the story.

Emphasizing data inferences and discussing results is crucial. Also, leveraging programming software like R and Python provides extensive flexibility to generate high-quality graphs, enhancing the overall standards of presentation.

I remember once presenting one graph at a conference and my audience completely ignoring it. Trend lines errors dots bla bla bla. I had forgotten about the graph, and re-graphed it using a much simpler visualization at another presentation... and I got more feedback! The first one was too wordy and complicated even for experts! The less people have to process, the better!

It is important to cover the various technical/methodological aspects of the project (eg: with dedicated, detailed yet clear slides, for the Q&A for instance). It is important to mention the limitations of the #study, as well as potential cognitive #biases at play, the concrete #results that were obtained, as well as short-term and midterm #perspectives