When Practice Became More Varied: An Early Observation from Vocora
A first-person note on variety, perceived mistakes, and the experience of practice
Early observation · one learner
I am both the developer of Vocora and the learner describing this experience. Mixed practice felt less repetitive, and I had the impression of making fewer mistakes and feeling less stress. I have not calculated a matched accuracy comparison or tested whether learning improved.
Earlier, much of my practice involved hearing vocabulary and writing its spelling. Other activities, such as grammar and speaking practice, were approached separately. A long run could therefore mean repeating the same kind of response many times.
I later brought different activities into a central Leitner practice flow. A session could now contain a combination of vocabulary, grammar, reading, and writing activities. I also introduced more variety in how a question was presented and answered, including several formats within grammar practice itself.
The most immediate difference was how the session felt. I was no longer doing the same thing throughout. I also had the impression that, for the same number of slides, I made fewer mistakes. Previously, repeated mistakes could make me feel stressed; with fewer mistakes apparent to me, that experience felt less stressful.
Table of contents
- Two kinds of variety changed together
- What the central practice flow changed
- What I noticed, and what remains unmeasured
- How interleaving research relates to this experience
- Why fewer apparent mistakes could have several explanations
- What would make the comparison informative
- Relationship to the other observations
- References
- Acknowledgement of AI Use
Two kinds of variety changed together
The first change concerned what I practiced. Spelling, word meaning, grammatical forms, and activities involving reading or writing could appear within one practice flow, instead of remaining in separate runs. A session contained a selection from the available activities; it did not necessarily include every category.
The second change concerned how I responded. Even when the target remained grammar, the interaction could change:
| Grammar exercise format | What the learner does |
|---|---|
| Cloze or gap filling | Completes a missing part of a sentence. |
| Multiple choice | Selects an answer from supplied alternatives. |
| Error identification and correction | Reads a text, finds a problem, and writes a correction. |
| Shuffled-word sentence assembly | Arranges supplied words into an appropriate sentence. |
This table describes the kinds of exercises in my account, not a fixed four-step sequence. It does not mean that every grammar target appeared in every format or that each session contained equal numbers of them.
These two dimensions are easy to confuse. Switching from spelling to grammar changes the content or skill being practiced. Switching from a grammar gap-fill to sentence assembly changes the response format, potentially while keeping a related grammatical target. My experience changed along both dimensions, so it does not isolate the contribution of either one.
What the central practice flow changed
The central Leitner flow provided a shared place for selecting and mixing activities according to the application’s rules. The relevant experience was moving between different kinds of work within a session. I am not reporting an unrestricted random shuffle of the entire learning library, an equal quota for every skill, or a measured comparison of scheduling algorithms.
There are also two different questions about order. One concerns when material returns for review. Another concerns which activities appear next to one another within a session. Bringing them into one system does not make the effects of review timing and within-session variety interchangeable.
This note focuses on the second question and on the experience of a mixed session. The earlier spelling-retention note concerned review routines and a later spelling check. Its reported figures do not measure the effect of mixing skills or formats.
What I noticed, and what remains unmeasured
The clearest observation was reduced monotony. Moving to a different kind of question interrupted the feeling of repeating one action continuously. I experienced the practice as more engaging and had a stronger sense that I was learning.
I also perceived fewer mistakes. My informal comparison was between runs containing the same number of slides, rather than between a short mixed session and a much longer single-format session. However, the same slide count does not establish equivalent practice. The questions could differ in difficulty, familiarity, time needed, or the amount of help built into the response format.
I have not calculated first-attempt accuracy for comparable items before and after the change. There is no documented error reduction, effect size, or fixed observation period in this account. The statement that mistakes seemed fewer should remain a statement about my impression until the underlying opportunities and outcomes are compared.
The stress observation has the same boundary. Repeated mistakes had felt discouraging to me, and the mixed practice felt less stressful when I perceived fewer errors. I did not collect stress ratings or physiological measures. The account cannot determine whether variety changed stress, whether different questions were easier, or whether feeling more comfortable changed how I noticed mistakes.
How interleaving research relates to this experience
Research on interleaving examines the order of practice, often by alternating related categories or problem types instead of completing a block of one type before moving to another. This is relevant to some aspects of the design, but mixing unrelated skills and changing an interface format do not automatically reproduce a studied interleaving condition.
Nakata and Suzuki (2019) examined blocked, interleaved, and initially blocked then interleaved practice of five English grammatical structures among Japanese learners. Interleaving produced more incorrect responses during training, yet better performance than blocking on a grammaticality judgment test one week later. Their design also changed the spacing between repetitions of each grammatical structure, so it did not isolate interleaving from spacing. Their study makes the distinction between practice errors and later retention especially important here. It does not support using my impression of fewer immediate mistakes as proof of an interleaving benefit.
Rogers and Monsell (1995) found performance costs when participants switched between simple classification tasks, even when the switches were predictable. Those laboratory tasks differ from language practice, but the result is a useful counterweight to the idea that more switching must always be better. A design can feel pleasantly varied while also imposing a cost when the learner changes task.
Kang, McDermott, and Roediger (2007) compared multiple-choice and short-answer practice on studied material. Their delayed results depended on both test format and corrective feedback. This supports treating format and feedback as separate parts of a comparison; it does not show that rotating formats improves learning or establish a preferred format for Vocora. An enjoyable session, a higher immediate score, and better later learning still require separate evidence. In my case, even the immediate score difference has not yet been established.
These sources support a careful investigation of practice order and outcomes. None tested Vocora’s mixture of content, response formats, scheduling rules, and feedback.
Why fewer apparent mistakes could have several explanations
Changing the mix may change the kinds of errors that are possible. Selecting a grammatical answer from alternatives, arranging visible words, and producing an answer without those choices place different demands on a learner. A correct response in one format should not be treated as interchangeable evidence of unaided production in another.
The mix might also contain a different proportion of familiar or difficult material. If a session includes fewer of the items I find hardest, its overall error rate could fall even if my performance within each kind of task remains unchanged. Keeping the number of slides constant does not remove that possibility.
Repeated exposure, changes in session size, feedback, and growing familiarity with the interface could contribute as well. I developed the changes and expected them to help, which creates another possible influence on my interpretation. Without a comparison that separates these elements, I cannot assign the experience to interleaving, format variety, or the central scheduler alone.
Reading a sentence to find a grammatical error is also a particular activity. Success on that exercise does not demonstrate a general improvement in reading comprehension, just as arranging supplied words does not establish better free writing or speaking.
What would make the comparison informative
When content, difficulty, practice opportunities, and feedback are comparable, does mixing skills or response formats change the experience of practice, first-attempt accuracy, or later learning?
The two forms of variety should be examined separately. One comparison could keep the content and response formats fixed while changing their order from blocked to mixed. Changing order may also change the gaps between repetitions of the same target. Unless those gaps are controlled, the comparison would estimate the combined effects of mixing and spacing. Another could keep the grammatical targets comparable while varying the response formats. A study of the complete product experience could compare the whole routines, but would need to report that several features differ together.
Useful records would include the target skill, item, response format, prior exposure, presentation order, review interval, time spent, and feedback received. First-attempt errors and success after a retry should be counted separately. Accuracy should be examined within comparable skills and formats as well as across the whole session, so that a change in composition is not mistaken for an improvement on the same work.
Brief ratings could assess monotony, engagement, and stress without treating them as learning measures. Delayed assessments should specify whether they test previously practiced items, new examples of the same targets, or broader transfer, using comparable test conditions across groups. Both benefits and costs of switching would remain possible outcomes.
Relationship to the other observations
My session-boundaries note concerns dividing the same daily workload into smaller, learner-selected sessions. Here, the question is what happens inside those sessions when the content and interactions vary.
Both arose within my ongoing use of Vocora. The changes may overlap in time, and the accounts do not isolate independent effects. The earlier notes on attention, grammar, and reflection also cannot be combined with this one into a single improvement estimate.
The experience I want to preserve is that mixed practice felt less repetitive, mistakes seemed fewer, and the repeated-error stress I had experienced felt less prominent. Whether those impressions correspond to improved performance or more durable learning remains an open question.
References
- Nakata, T., & Suzuki, Y. (2019). Mixing grammar exercises facilitates long-term retention: Effects of blocking, interleaving, and increasing practice. The Modern Language Journal, 103(3), 629–647.
- Rogers, R. D., & Monsell, S. (1995). Costs of a predictable switch between simple cognitive tasks. Journal of Experimental Psychology: General, 124(2), 207–231.
- Kang, S. H. K., McDermott, K. B., & Roediger, H. L., III. (2007). Test format and corrective feedback modify the effect of testing on long-term retention. European Journal of Cognitive Psychology, 19(4–5), 528–558.
Acknowledgement of AI Use
ChatGPT (OpenAI) assisted with drafting, organizing, and editing this note. The author supplied the underlying observations and retains responsibility for the account and its interpretation.